vcn_v1_0.c 32.6 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 25 26 27
/*
 * Copyright 2016 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/firmware.h>
#include <drm/drmP.h>
#include "amdgpu.h"
#include "amdgpu_vcn.h"
28
#include "soc15.h"
29 30 31
#include "soc15d.h"
#include "soc15_common.h"

32 33
#include "vcn/vcn_1_0_offset.h"
#include "vcn/vcn_1_0_sh_mask.h"
34
#include "hdp/hdp_4_0_offset.h"
35 36
#include "mmhub/mmhub_9_1_offset.h"
#include "mmhub/mmhub_9_1_sh_mask.h"
37 38 39

static int vcn_v1_0_start(struct amdgpu_device *adev);
static int vcn_v1_0_stop(struct amdgpu_device *adev);
40
static void vcn_v1_0_set_dec_ring_funcs(struct amdgpu_device *adev);
41
static void vcn_v1_0_set_enc_ring_funcs(struct amdgpu_device *adev);
L
Leo Liu 已提交
42
static void vcn_v1_0_set_irq_funcs(struct amdgpu_device *adev);
43 44 45 46 47 48 49 50 51 52

/**
 * vcn_v1_0_early_init - set function pointers
 *
 * @handle: amdgpu_device pointer
 *
 * Set ring and irq function pointers
 */
static int vcn_v1_0_early_init(void *handle)
{
53 54
	struct amdgpu_device *adev = (struct amdgpu_device *)handle;

L
Leo Liu 已提交
55 56
	adev->vcn.num_enc_rings = 2;

57
	vcn_v1_0_set_dec_ring_funcs(adev);
58
	vcn_v1_0_set_enc_ring_funcs(adev);
L
Leo Liu 已提交
59
	vcn_v1_0_set_irq_funcs(adev);
60

61 62 63 64 65 66 67 68 69 70 71 72
	return 0;
}

/**
 * vcn_v1_0_sw_init - sw init for VCN block
 *
 * @handle: amdgpu_device pointer
 *
 * Load firmware and sw initialization
 */
static int vcn_v1_0_sw_init(void *handle)
{
73
	struct amdgpu_ring *ring;
L
Leo Liu 已提交
74
	int i, r;
75 76
	struct amdgpu_device *adev = (struct amdgpu_device *)handle;

L
Leo Liu 已提交
77
	/* VCN DEC TRAP */
78 79 80 81
	r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_VCN, 124, &adev->vcn.irq);
	if (r)
		return r;

L
Leo Liu 已提交
82 83 84 85 86 87 88 89
	/* VCN ENC TRAP */
	for (i = 0; i < adev->vcn.num_enc_rings; ++i) {
		r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_VCN, i + 119,
					&adev->vcn.irq);
		if (r)
			return r;
	}

90 91 92 93 94 95 96 97
	r = amdgpu_vcn_sw_init(adev);
	if (r)
		return r;

	r = amdgpu_vcn_resume(adev);
	if (r)
		return r;

98 99 100
	ring = &adev->vcn.ring_dec;
	sprintf(ring->name, "vcn_dec");
	r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.irq, 0);
L
Leo Liu 已提交
101 102 103 104 105 106 107 108 109 110
	if (r)
		return r;

	for (i = 0; i < adev->vcn.num_enc_rings; ++i) {
		ring = &adev->vcn.ring_enc[i];
		sprintf(ring->name, "vcn_enc%d", i);
		r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.irq, 0);
		if (r)
			return r;
	}
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
	return r;
}

/**
 * vcn_v1_0_sw_fini - sw fini for VCN block
 *
 * @handle: amdgpu_device pointer
 *
 * VCN suspend and free up sw allocation
 */
static int vcn_v1_0_sw_fini(void *handle)
{
	int r;
	struct amdgpu_device *adev = (struct amdgpu_device *)handle;

	r = amdgpu_vcn_suspend(adev);
	if (r)
		return r;

	r = amdgpu_vcn_sw_fini(adev);

	return r;
}

/**
 * vcn_v1_0_hw_init - start and test VCN block
 *
 * @handle: amdgpu_device pointer
 *
 * Initialize the hardware, boot up the VCPU and do some testing
 */
static int vcn_v1_0_hw_init(void *handle)
{
	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
	struct amdgpu_ring *ring = &adev->vcn.ring_dec;
147
	int i, r;
148 149 150 151 152 153 154 155 156 157 158 159

	r = vcn_v1_0_start(adev);
	if (r)
		goto done;

	ring->ready = true;
	r = amdgpu_ring_test_ring(ring);
	if (r) {
		ring->ready = false;
		goto done;
	}

160 161 162 163 164 165 166 167 168 169
	for (i = 0; i < adev->vcn.num_enc_rings; ++i) {
		ring = &adev->vcn.ring_enc[i];
		ring->ready = true;
		r = amdgpu_ring_test_ring(ring);
		if (r) {
			ring->ready = false;
			goto done;
		}
	}

170 171
done:
	if (!r)
172
		DRM_INFO("VCN decode and encode initialized successfully.\n");
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

	return r;
}

/**
 * vcn_v1_0_hw_fini - stop the hardware block
 *
 * @handle: amdgpu_device pointer
 *
 * Stop the VCN block, mark ring as not ready any more
 */
static int vcn_v1_0_hw_fini(void *handle)
{
	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
	struct amdgpu_ring *ring = &adev->vcn.ring_dec;
	int r;

	r = vcn_v1_0_stop(adev);
	if (r)
		return r;

	ring->ready = false;

	return 0;
}

/**
 * vcn_v1_0_suspend - suspend VCN block
 *
 * @handle: amdgpu_device pointer
 *
 * HW fini and suspend VCN block
 */
static int vcn_v1_0_suspend(void *handle)
{
	int r;
	struct amdgpu_device *adev = (struct amdgpu_device *)handle;

	r = vcn_v1_0_hw_fini(adev);
	if (r)
		return r;

	r = amdgpu_vcn_suspend(adev);

	return r;
}

/**
 * vcn_v1_0_resume - resume VCN block
 *
 * @handle: amdgpu_device pointer
 *
 * Resume firmware and hw init VCN block
 */
static int vcn_v1_0_resume(void *handle)
{
	int r;
	struct amdgpu_device *adev = (struct amdgpu_device *)handle;

	r = amdgpu_vcn_resume(adev);
	if (r)
		return r;

	r = vcn_v1_0_hw_init(adev);

	return r;
}

/**
 * vcn_v1_0_mc_resume - memory controller programming
 *
 * @adev: amdgpu_device pointer
 *
 * Let the VCN memory controller know it's offsets
 */
static void vcn_v1_0_mc_resume(struct amdgpu_device *adev)
{
250
	uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw->size + 4);
251

252
	WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
253
			lower_32_bits(adev->vcn.gpu_addr));
254
	WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
255
			upper_32_bits(adev->vcn.gpu_addr));
256
	WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_OFFSET0,
257
				AMDGPU_UVD_FIRMWARE_OFFSET >> 3);
258
	WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_SIZE0, size);
259

260
	WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW,
261
			lower_32_bits(adev->vcn.gpu_addr + size));
262
	WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH,
263
			upper_32_bits(adev->vcn.gpu_addr + size));
264 265
	WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_OFFSET1, 0);
	WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_SIZE1, AMDGPU_VCN_HEAP_SIZE);
266

267
	WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW,
268
			lower_32_bits(adev->vcn.gpu_addr + size + AMDGPU_VCN_HEAP_SIZE));
269
	WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH,
270
			upper_32_bits(adev->vcn.gpu_addr + size + AMDGPU_VCN_HEAP_SIZE));
271 272
	WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_OFFSET2, 0);
	WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_SIZE2,
273
			AMDGPU_VCN_STACK_SIZE + (AMDGPU_VCN_SESSION_SIZE * 40));
274

275
	WREG32_SOC15(UVD, 0, mmUVD_UDEC_ADDR_CONFIG,
276
			adev->gfx.config.gb_addr_config);
277
	WREG32_SOC15(UVD, 0, mmUVD_UDEC_DB_ADDR_CONFIG,
278
			adev->gfx.config.gb_addr_config);
279
	WREG32_SOC15(UVD, 0, mmUVD_UDEC_DBW_ADDR_CONFIG,
280 281 282
			adev->gfx.config.gb_addr_config);
}

H
Huang Rui 已提交
283 284 285 286 287 288 289 290 291 292 293 294 295
/**
 * vcn_v1_0_disable_clock_gating - disable VCN clock gating
 *
 * @adev: amdgpu_device pointer
 * @sw: enable SW clock gating
 *
 * Disable clock gating for VCN block
 */
static void vcn_v1_0_disable_clock_gating(struct amdgpu_device *adev, bool sw)
{
	uint32_t data;

	/* JPEG disable CGC */
296
	data = RREG32_SOC15(VCN, 0, mmJPEG_CGC_CTRL);
H
Huang Rui 已提交
297 298 299 300 301 302 303 304

	if (sw)
		data |= 1 << JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
	else
		data &= ~JPEG_CGC_CTRL__DYN_CLOCK_MODE_MASK;

	data |= 1 << JPEG_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
	data |= 4 << JPEG_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
305
	WREG32_SOC15(VCN, 0, mmJPEG_CGC_CTRL, data);
H
Huang Rui 已提交
306

307
	data = RREG32_SOC15(VCN, 0, mmJPEG_CGC_GATE);
H
Huang Rui 已提交
308
	data &= ~(JPEG_CGC_GATE__JPEG_MASK | JPEG_CGC_GATE__JPEG2_MASK);
309
	WREG32_SOC15(VCN, 0, mmJPEG_CGC_GATE, data);
H
Huang Rui 已提交
310 311

	/* UVD disable CGC */
312
	data = RREG32_SOC15(VCN, 0, mmUVD_CGC_CTRL);
H
Huang Rui 已提交
313 314 315 316 317 318 319
	if (sw)
		data |= 1 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
	else
		data &= ~ UVD_CGC_CTRL__DYN_CLOCK_MODE_MASK;

	data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
	data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
320
	WREG32_SOC15(VCN, 0, mmUVD_CGC_CTRL, data);
H
Huang Rui 已提交
321

322
	data = RREG32_SOC15(VCN, 0, mmUVD_CGC_GATE);
H
Huang Rui 已提交
323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342
	data &= ~(UVD_CGC_GATE__SYS_MASK
		| UVD_CGC_GATE__UDEC_MASK
		| UVD_CGC_GATE__MPEG2_MASK
		| UVD_CGC_GATE__REGS_MASK
		| UVD_CGC_GATE__RBC_MASK
		| UVD_CGC_GATE__LMI_MC_MASK
		| UVD_CGC_GATE__LMI_UMC_MASK
		| UVD_CGC_GATE__IDCT_MASK
		| UVD_CGC_GATE__MPRD_MASK
		| UVD_CGC_GATE__MPC_MASK
		| UVD_CGC_GATE__LBSI_MASK
		| UVD_CGC_GATE__LRBBM_MASK
		| UVD_CGC_GATE__UDEC_RE_MASK
		| UVD_CGC_GATE__UDEC_CM_MASK
		| UVD_CGC_GATE__UDEC_IT_MASK
		| UVD_CGC_GATE__UDEC_DB_MASK
		| UVD_CGC_GATE__UDEC_MP_MASK
		| UVD_CGC_GATE__WCB_MASK
		| UVD_CGC_GATE__VCPU_MASK
		| UVD_CGC_GATE__SCPU_MASK);
343
	WREG32_SOC15(VCN, 0, mmUVD_CGC_GATE, data);
H
Huang Rui 已提交
344

345
	data = RREG32_SOC15(VCN, 0, mmUVD_CGC_CTRL);
H
Huang Rui 已提交
346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365
	data &= ~(UVD_CGC_CTRL__UDEC_RE_MODE_MASK
		| UVD_CGC_CTRL__UDEC_CM_MODE_MASK
		| UVD_CGC_CTRL__UDEC_IT_MODE_MASK
		| UVD_CGC_CTRL__UDEC_DB_MODE_MASK
		| UVD_CGC_CTRL__UDEC_MP_MODE_MASK
		| UVD_CGC_CTRL__SYS_MODE_MASK
		| UVD_CGC_CTRL__UDEC_MODE_MASK
		| UVD_CGC_CTRL__MPEG2_MODE_MASK
		| UVD_CGC_CTRL__REGS_MODE_MASK
		| UVD_CGC_CTRL__RBC_MODE_MASK
		| UVD_CGC_CTRL__LMI_MC_MODE_MASK
		| UVD_CGC_CTRL__LMI_UMC_MODE_MASK
		| UVD_CGC_CTRL__IDCT_MODE_MASK
		| UVD_CGC_CTRL__MPRD_MODE_MASK
		| UVD_CGC_CTRL__MPC_MODE_MASK
		| UVD_CGC_CTRL__LBSI_MODE_MASK
		| UVD_CGC_CTRL__LRBBM_MODE_MASK
		| UVD_CGC_CTRL__WCB_MODE_MASK
		| UVD_CGC_CTRL__VCPU_MODE_MASK
		| UVD_CGC_CTRL__SCPU_MODE_MASK);
366
	WREG32_SOC15(VCN, 0, mmUVD_CGC_CTRL, data);
H
Huang Rui 已提交
367 368

	/* turn on */
369
	data = RREG32_SOC15(VCN, 0, mmUVD_SUVD_CGC_GATE);
H
Huang Rui 已提交
370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393
	data |= (UVD_SUVD_CGC_GATE__SRE_MASK
		| UVD_SUVD_CGC_GATE__SIT_MASK
		| UVD_SUVD_CGC_GATE__SMP_MASK
		| UVD_SUVD_CGC_GATE__SCM_MASK
		| UVD_SUVD_CGC_GATE__SDB_MASK
		| UVD_SUVD_CGC_GATE__SRE_H264_MASK
		| UVD_SUVD_CGC_GATE__SRE_HEVC_MASK
		| UVD_SUVD_CGC_GATE__SIT_H264_MASK
		| UVD_SUVD_CGC_GATE__SIT_HEVC_MASK
		| UVD_SUVD_CGC_GATE__SCM_H264_MASK
		| UVD_SUVD_CGC_GATE__SCM_HEVC_MASK
		| UVD_SUVD_CGC_GATE__SDB_H264_MASK
		| UVD_SUVD_CGC_GATE__SDB_HEVC_MASK
		| UVD_SUVD_CGC_GATE__SCLR_MASK
		| UVD_SUVD_CGC_GATE__UVD_SC_MASK
		| UVD_SUVD_CGC_GATE__ENT_MASK
		| UVD_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK
		| UVD_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK
		| UVD_SUVD_CGC_GATE__SITE_MASK
		| UVD_SUVD_CGC_GATE__SRE_VP9_MASK
		| UVD_SUVD_CGC_GATE__SCM_VP9_MASK
		| UVD_SUVD_CGC_GATE__SIT_VP9_DEC_MASK
		| UVD_SUVD_CGC_GATE__SDB_VP9_MASK
		| UVD_SUVD_CGC_GATE__IME_HEVC_MASK);
394
	WREG32_SOC15(VCN, 0, mmUVD_SUVD_CGC_GATE, data);
H
Huang Rui 已提交
395

396
	data = RREG32_SOC15(VCN, 0, mmUVD_SUVD_CGC_CTRL);
H
Huang Rui 已提交
397 398 399 400 401 402 403 404 405 406
	data &= ~(UVD_SUVD_CGC_CTRL__SRE_MODE_MASK
		| UVD_SUVD_CGC_CTRL__SIT_MODE_MASK
		| UVD_SUVD_CGC_CTRL__SMP_MODE_MASK
		| UVD_SUVD_CGC_CTRL__SCM_MODE_MASK
		| UVD_SUVD_CGC_CTRL__SDB_MODE_MASK
		| UVD_SUVD_CGC_CTRL__SCLR_MODE_MASK
		| UVD_SUVD_CGC_CTRL__UVD_SC_MODE_MASK
		| UVD_SUVD_CGC_CTRL__ENT_MODE_MASK
		| UVD_SUVD_CGC_CTRL__IME_MODE_MASK
		| UVD_SUVD_CGC_CTRL__SITE_MODE_MASK);
407
	WREG32_SOC15(VCN, 0, mmUVD_SUVD_CGC_CTRL, data);
H
Huang Rui 已提交
408 409 410 411 412 413 414 415 416 417 418 419 420 421 422
}

/**
 * vcn_v1_0_enable_clock_gating - enable VCN clock gating
 *
 * @adev: amdgpu_device pointer
 * @sw: enable SW clock gating
 *
 * Enable clock gating for VCN block
 */
static void vcn_v1_0_enable_clock_gating(struct amdgpu_device *adev, bool sw)
{
	uint32_t data = 0;

	/* enable JPEG CGC */
423
	data = RREG32_SOC15(VCN, 0, mmJPEG_CGC_CTRL);
H
Huang Rui 已提交
424 425 426 427 428 429
	if (sw)
		data |= 1 << JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
	else
		data |= 0 << JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
	data |= 1 << JPEG_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
	data |= 4 << JPEG_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
430
	WREG32_SOC15(VCN, 0, mmJPEG_CGC_CTRL, data);
H
Huang Rui 已提交
431

432
	data = RREG32_SOC15(VCN, 0, mmJPEG_CGC_GATE);
H
Huang Rui 已提交
433
	data |= (JPEG_CGC_GATE__JPEG_MASK | JPEG_CGC_GATE__JPEG2_MASK);
434
	WREG32_SOC15(VCN, 0, mmJPEG_CGC_GATE, data);
H
Huang Rui 已提交
435 436

	/* enable UVD CGC */
437
	data = RREG32_SOC15(VCN, 0, mmUVD_CGC_CTRL);
H
Huang Rui 已提交
438 439 440 441 442 443
	if (sw)
		data |= 1 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
	else
		data |= 0 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
	data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
	data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
444
	WREG32_SOC15(VCN, 0, mmUVD_CGC_CTRL, data);
H
Huang Rui 已提交
445

446
	data = RREG32_SOC15(VCN, 0, mmUVD_CGC_CTRL);
H
Huang Rui 已提交
447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466
	data |= (UVD_CGC_CTRL__UDEC_RE_MODE_MASK
		| UVD_CGC_CTRL__UDEC_CM_MODE_MASK
		| UVD_CGC_CTRL__UDEC_IT_MODE_MASK
		| UVD_CGC_CTRL__UDEC_DB_MODE_MASK
		| UVD_CGC_CTRL__UDEC_MP_MODE_MASK
		| UVD_CGC_CTRL__SYS_MODE_MASK
		| UVD_CGC_CTRL__UDEC_MODE_MASK
		| UVD_CGC_CTRL__MPEG2_MODE_MASK
		| UVD_CGC_CTRL__REGS_MODE_MASK
		| UVD_CGC_CTRL__RBC_MODE_MASK
		| UVD_CGC_CTRL__LMI_MC_MODE_MASK
		| UVD_CGC_CTRL__LMI_UMC_MODE_MASK
		| UVD_CGC_CTRL__IDCT_MODE_MASK
		| UVD_CGC_CTRL__MPRD_MODE_MASK
		| UVD_CGC_CTRL__MPC_MODE_MASK
		| UVD_CGC_CTRL__LBSI_MODE_MASK
		| UVD_CGC_CTRL__LRBBM_MODE_MASK
		| UVD_CGC_CTRL__WCB_MODE_MASK
		| UVD_CGC_CTRL__VCPU_MODE_MASK
		| UVD_CGC_CTRL__SCPU_MODE_MASK);
467
	WREG32_SOC15(VCN, 0, mmUVD_CGC_CTRL, data);
H
Huang Rui 已提交
468

469
	data = RREG32_SOC15(VCN, 0, mmUVD_SUVD_CGC_CTRL);
H
Huang Rui 已提交
470 471 472 473 474 475 476 477 478 479
	data |= (UVD_SUVD_CGC_CTRL__SRE_MODE_MASK
		| UVD_SUVD_CGC_CTRL__SIT_MODE_MASK
		| UVD_SUVD_CGC_CTRL__SMP_MODE_MASK
		| UVD_SUVD_CGC_CTRL__SCM_MODE_MASK
		| UVD_SUVD_CGC_CTRL__SDB_MODE_MASK
		| UVD_SUVD_CGC_CTRL__SCLR_MODE_MASK
		| UVD_SUVD_CGC_CTRL__UVD_SC_MODE_MASK
		| UVD_SUVD_CGC_CTRL__ENT_MODE_MASK
		| UVD_SUVD_CGC_CTRL__IME_MODE_MASK
		| UVD_SUVD_CGC_CTRL__SITE_MODE_MASK);
480
	WREG32_SOC15(VCN, 0, mmUVD_SUVD_CGC_CTRL, data);
H
Huang Rui 已提交
481 482
}

483 484 485 486 487 488 489 490 491
/**
 * vcn_v1_0_start - start VCN block
 *
 * @adev: amdgpu_device pointer
 *
 * Setup and start the VCN block
 */
static int vcn_v1_0_start(struct amdgpu_device *adev)
{
492 493
	struct amdgpu_ring *ring = &adev->vcn.ring_dec;
	uint32_t rb_bufsz, tmp;
494 495 496 497 498 499 500 501 502
	uint32_t lmi_swap_cntl;
	int i, j, r;

	/* disable byte swapping */
	lmi_swap_cntl = 0;

	vcn_v1_0_mc_resume(adev);

	/* disable clock gating */
503
	vcn_v1_0_disable_clock_gating(adev, true);
504 505 506 507 508 509 510 511 512 513 514 515

	/* disable interupt */
	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_MASTINT_EN), 0,
			~UVD_MASTINT_EN__VCPU_EN_MASK);

	/* stall UMC and register bus before resetting VCPU */
	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_CTRL2),
			UVD_LMI_CTRL2__STALL_ARB_UMC_MASK,
			~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);
	mdelay(1);

	/* put LMI, VCPU, RBC etc... into reset */
516
	WREG32_SOC15(UVD, 0, mmUVD_SOFT_RESET,
517 518 519 520 521 522 523 524 525 526 527
		UVD_SOFT_RESET__LMI_SOFT_RESET_MASK |
		UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK |
		UVD_SOFT_RESET__LBSI_SOFT_RESET_MASK |
		UVD_SOFT_RESET__RBC_SOFT_RESET_MASK |
		UVD_SOFT_RESET__CSM_SOFT_RESET_MASK |
		UVD_SOFT_RESET__CXW_SOFT_RESET_MASK |
		UVD_SOFT_RESET__TAP_SOFT_RESET_MASK |
		UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK);
	mdelay(5);

	/* initialize VCN memory controller */
528
	WREG32_SOC15(UVD, 0, mmUVD_LMI_CTRL,
529 530 531 532 533 534 535 536 537 538 539
		(0x40 << UVD_LMI_CTRL__WRITE_CLEAN_TIMER__SHIFT) |
		UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK |
		UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
		UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK |
		UVD_LMI_CTRL__REQ_MODE_MASK |
		0x00100000L);

#ifdef __BIG_ENDIAN
	/* swap (8 in 32) RB and IB */
	lmi_swap_cntl = 0xa;
#endif
540
	WREG32_SOC15(UVD, 0, mmUVD_LMI_SWAP_CNTL, lmi_swap_cntl);
541

542 543 544 545 546 547
	WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_MUXA0, 0x40c2040);
	WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_MUXA1, 0x0);
	WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_MUXB0, 0x40c2040);
	WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_MUXB1, 0x0);
	WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_ALU, 0);
	WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_MUX, 0x88);
548 549

	/* take all subblocks out of reset, except VCPU */
550
	WREG32_SOC15(UVD, 0, mmUVD_SOFT_RESET,
551 552 553 554
			UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK);
	mdelay(5);

	/* enable VCPU clock */
555
	WREG32_SOC15(UVD, 0, mmUVD_VCPU_CNTL,
556 557 558 559 560 561 562
			UVD_VCPU_CNTL__CLK_EN_MASK);

	/* enable UMC */
	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_CTRL2), 0,
			~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);

	/* boot up the VCPU */
563
	WREG32_SOC15(UVD, 0, mmUVD_SOFT_RESET, 0);
564 565 566 567 568 569
	mdelay(10);

	for (i = 0; i < 10; ++i) {
		uint32_t status;

		for (j = 0; j < 100; ++j) {
570
			status = RREG32_SOC15(UVD, 0, mmUVD_STATUS);
571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602
			if (status & 2)
				break;
			mdelay(10);
		}
		r = 0;
		if (status & 2)
			break;

		DRM_ERROR("VCN decode not responding, trying to reset the VCPU!!!\n");
		WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_SOFT_RESET),
				UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK,
				~UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK);
		mdelay(10);
		WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_SOFT_RESET), 0,
				~UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK);
		mdelay(10);
		r = -1;
	}

	if (r) {
		DRM_ERROR("VCN decode not responding, giving up!!!\n");
		return r;
	}
	/* enable master interrupt */
	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_MASTINT_EN),
		(UVD_MASTINT_EN__VCPU_EN_MASK|UVD_MASTINT_EN__SYS_EN_MASK),
		~(UVD_MASTINT_EN__VCPU_EN_MASK|UVD_MASTINT_EN__SYS_EN_MASK));

	/* clear the bit 4 of VCN_STATUS */
	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_STATUS), 0,
			~(2 << UVD_STATUS__VCPU_REPORT__SHIFT));

603 604 605 606 607 608 609 610
	/* force RBC into idle state */
	rb_bufsz = order_base_2(ring->ring_size);
	tmp = REG_SET_FIELD(0, UVD_RBC_RB_CNTL, RB_BUFSZ, rb_bufsz);
	tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_BLKSZ, 1);
	tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_FETCH, 1);
	tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_WPTR_POLL_EN, 0);
	tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_UPDATE, 1);
	tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_RPTR_WR_EN, 1);
611
	WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_CNTL, tmp);
612 613

	/* set the write pointer delay */
614
	WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR_CNTL, 0);
615 616

	/* set the wb address */
617
	WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_RPTR_ADDR,
618 619 620
			(upper_32_bits(ring->gpu_addr) >> 2));

	/* programm the RB_BASE for ring buffer */
621
	WREG32_SOC15(UVD, 0, mmUVD_LMI_RBC_RB_64BIT_BAR_LOW,
622
			lower_32_bits(ring->gpu_addr));
623
	WREG32_SOC15(UVD, 0, mmUVD_LMI_RBC_RB_64BIT_BAR_HIGH,
624 625 626
			upper_32_bits(ring->gpu_addr));

	/* Initialize the ring buffer's read and write pointers */
627
	WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_RPTR, 0);
628

629 630
	ring->wptr = RREG32_SOC15(UVD, 0, mmUVD_RBC_RB_RPTR);
	WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR,
631 632 633 634 635
			lower_32_bits(ring->wptr));

	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_RBC_RB_CNTL), 0,
			~UVD_RBC_RB_CNTL__RB_NO_FETCH_MASK);

L
Leo Liu 已提交
636
	ring = &adev->vcn.ring_enc[0];
637 638 639 640 641
	WREG32_SOC15(UVD, 0, mmUVD_RB_RPTR, lower_32_bits(ring->wptr));
	WREG32_SOC15(UVD, 0, mmUVD_RB_WPTR, lower_32_bits(ring->wptr));
	WREG32_SOC15(UVD, 0, mmUVD_RB_BASE_LO, ring->gpu_addr);
	WREG32_SOC15(UVD, 0, mmUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr));
	WREG32_SOC15(UVD, 0, mmUVD_RB_SIZE, ring->ring_size / 4);
L
Leo Liu 已提交
642 643

	ring = &adev->vcn.ring_enc[1];
644 645 646 647 648
	WREG32_SOC15(UVD, 0, mmUVD_RB_RPTR2, lower_32_bits(ring->wptr));
	WREG32_SOC15(UVD, 0, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr));
	WREG32_SOC15(UVD, 0, mmUVD_RB_BASE_LO2, ring->gpu_addr);
	WREG32_SOC15(UVD, 0, mmUVD_RB_BASE_HI2, upper_32_bits(ring->gpu_addr));
	WREG32_SOC15(UVD, 0, mmUVD_RB_SIZE2, ring->ring_size / 4);
L
Leo Liu 已提交
649

650 651 652 653 654 655 656 657 658 659 660 661
	return 0;
}

/**
 * vcn_v1_0_stop - stop VCN block
 *
 * @adev: amdgpu_device pointer
 *
 * stop the VCN block
 */
static int vcn_v1_0_stop(struct amdgpu_device *adev)
{
662
	/* force RBC into idle state */
663
	WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_CNTL, 0x11010101);
664

665 666 667 668 669 670 671
	/* Stall UMC and register bus before resetting VCPU */
	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_CTRL2),
			UVD_LMI_CTRL2__STALL_ARB_UMC_MASK,
			~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);
	mdelay(1);

	/* put VCPU into reset */
672
	WREG32_SOC15(UVD, 0, mmUVD_SOFT_RESET,
673 674 675 676
			UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK);
	mdelay(5);

	/* disable VCPU clock */
677
	WREG32_SOC15(UVD, 0, mmUVD_VCPU_CNTL, 0x0);
678 679 680 681 682

	/* Unstall UMC and register bus */
	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_CTRL2), 0,
			~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);

H
Huang Rui 已提交
683
	/* enable clock gating */
684
	vcn_v1_0_enable_clock_gating(adev, true);
H
Huang Rui 已提交
685

686 687 688 689 690 691 692 693 694 695
	return 0;
}

static int vcn_v1_0_set_clockgating_state(void *handle,
					  enum amd_clockgating_state state)
{
	/* needed for driver unload*/
	return 0;
}

696 697 698 699 700 701 702 703 704 705 706
/**
 * vcn_v1_0_dec_ring_get_rptr - get read pointer
 *
 * @ring: amdgpu_ring pointer
 *
 * Returns the current hardware read pointer
 */
static uint64_t vcn_v1_0_dec_ring_get_rptr(struct amdgpu_ring *ring)
{
	struct amdgpu_device *adev = ring->adev;

707
	return RREG32_SOC15(UVD, 0, mmUVD_RBC_RB_RPTR);
708 709 710 711 712 713 714 715 716 717 718 719 720
}

/**
 * vcn_v1_0_dec_ring_get_wptr - get write pointer
 *
 * @ring: amdgpu_ring pointer
 *
 * Returns the current hardware write pointer
 */
static uint64_t vcn_v1_0_dec_ring_get_wptr(struct amdgpu_ring *ring)
{
	struct amdgpu_device *adev = ring->adev;

721
	return RREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR);
722 723 724 725 726 727 728 729 730 731 732 733 734
}

/**
 * vcn_v1_0_dec_ring_set_wptr - set write pointer
 *
 * @ring: amdgpu_ring pointer
 *
 * Commits the write pointer to the hardware
 */
static void vcn_v1_0_dec_ring_set_wptr(struct amdgpu_ring *ring)
{
	struct amdgpu_device *adev = ring->adev;

735
	WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR, lower_32_bits(ring->wptr));
736 737
}

738 739 740 741 742 743 744 745 746
/**
 * vcn_v1_0_dec_ring_insert_start - insert a start command
 *
 * @ring: amdgpu_ring pointer
 *
 * Write a start command to the ring.
 */
static void vcn_v1_0_dec_ring_insert_start(struct amdgpu_ring *ring)
{
747 748
	struct amdgpu_device *adev = ring->adev;

749 750 751 752 753
	amdgpu_ring_write(ring,
		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA0), 0));
	amdgpu_ring_write(ring, 0);
	amdgpu_ring_write(ring,
		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_CMD), 0));
754
	amdgpu_ring_write(ring, VCN_DEC_CMD_PACKET_START << 1);
755 756
}

757 758 759 760 761 762 763 764 765
/**
 * vcn_v1_0_dec_ring_insert_end - insert a end command
 *
 * @ring: amdgpu_ring pointer
 *
 * Write a end command to the ring.
 */
static void vcn_v1_0_dec_ring_insert_end(struct amdgpu_ring *ring)
{
766 767
	struct amdgpu_device *adev = ring->adev;

768 769
	amdgpu_ring_write(ring,
		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_CMD), 0));
770
	amdgpu_ring_write(ring, VCN_DEC_CMD_PACKET_END << 1);
771 772
}

773 774 775 776 777 778 779 780 781 782 783
/**
 * vcn_v1_0_dec_ring_emit_fence - emit an fence & trap command
 *
 * @ring: amdgpu_ring pointer
 * @fence: fence to emit
 *
 * Write a fence and a trap command to the ring.
 */
static void vcn_v1_0_dec_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq,
				     unsigned flags)
{
784 785
	struct amdgpu_device *adev = ring->adev;

786 787 788 789 790 791 792 793 794 795 796 797 798
	WARN_ON(flags & AMDGPU_FENCE_FLAG_64BIT);

	amdgpu_ring_write(ring,
		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_CONTEXT_ID), 0));
	amdgpu_ring_write(ring, seq);
	amdgpu_ring_write(ring,
		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA0), 0));
	amdgpu_ring_write(ring, addr & 0xffffffff);
	amdgpu_ring_write(ring,
		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA1), 0));
	amdgpu_ring_write(ring, upper_32_bits(addr) & 0xff);
	amdgpu_ring_write(ring,
		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_CMD), 0));
799
	amdgpu_ring_write(ring, VCN_DEC_CMD_FENCE << 1);
800 801 802 803 804 805 806 807 808

	amdgpu_ring_write(ring,
		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA0), 0));
	amdgpu_ring_write(ring, 0);
	amdgpu_ring_write(ring,
		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA1), 0));
	amdgpu_ring_write(ring, 0);
	amdgpu_ring_write(ring,
		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_CMD), 0));
809
	amdgpu_ring_write(ring, VCN_DEC_CMD_TRAP << 1);
810 811 812 813 814 815 816 817 818 819 820 821
}

/**
 * vcn_v1_0_dec_ring_emit_ib - execute indirect buffer
 *
 * @ring: amdgpu_ring pointer
 * @ib: indirect buffer to execute
 *
 * Write ring commands to execute the indirect buffer
 */
static void vcn_v1_0_dec_ring_emit_ib(struct amdgpu_ring *ring,
				  struct amdgpu_ib *ib,
822
				  unsigned vmid, bool ctx_switch)
823
{
824 825
	struct amdgpu_device *adev = ring->adev;

826 827
	amdgpu_ring_write(ring,
		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_RBC_IB_VMID), 0));
828
	amdgpu_ring_write(ring, vmid);
829 830 831 832 833 834 835 836 837 838 839 840 841 842 843

	amdgpu_ring_write(ring,
		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_RBC_IB_64BIT_BAR_LOW), 0));
	amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr));
	amdgpu_ring_write(ring,
		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_RBC_IB_64BIT_BAR_HIGH), 0));
	amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
	amdgpu_ring_write(ring,
		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_RBC_IB_SIZE), 0));
	amdgpu_ring_write(ring, ib->length_dw);
}

static void vcn_v1_0_dec_vm_reg_wait(struct amdgpu_ring *ring,
				uint32_t data0, uint32_t data1, uint32_t mask)
{
844 845
	struct amdgpu_device *adev = ring->adev;

846 847 848 849 850 851 852 853 854 855 856
	amdgpu_ring_write(ring,
		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA0), 0));
	amdgpu_ring_write(ring, data0);
	amdgpu_ring_write(ring,
		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA1), 0));
	amdgpu_ring_write(ring, data1);
	amdgpu_ring_write(ring,
		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GP_SCRATCH8), 0));
	amdgpu_ring_write(ring, mask);
	amdgpu_ring_write(ring,
		PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_CMD), 0));
857
	amdgpu_ring_write(ring, VCN_DEC_CMD_REG_READ_COND_WAIT << 1);
858 859 860
}

static void vcn_v1_0_dec_ring_emit_vm_flush(struct amdgpu_ring *ring,
861 862
					    unsigned vmid, unsigned pasid,
					    uint64_t pd_addr)
863 864 865
{
	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
	unsigned eng = ring->vm_inv_eng;
866
	uint32_t data0, data1, mask;
867

868
	pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
869

870
	/* wait for register write */
871
	data0 = (hub->ctx0_ptb_addr_lo32 + vmid * 2) << 2;
872 873 874 875 876 877
	data1 = lower_32_bits(pd_addr);
	mask = 0xffffffff;
	vcn_v1_0_dec_vm_reg_wait(ring, data0, data1, mask);

	/* wait for flush */
	data0 = (hub->vm_inv_eng0_ack + eng) << 2;
878 879
	data1 = 1 << vmid;
	mask =  1 << vmid;
880 881 882
	vcn_v1_0_dec_vm_reg_wait(ring, data0, data1, mask);
}

883 884 885 886 887 888 889 890 891 892 893 894
/**
 * vcn_v1_0_enc_ring_get_rptr - get enc read pointer
 *
 * @ring: amdgpu_ring pointer
 *
 * Returns the current hardware enc read pointer
 */
static uint64_t vcn_v1_0_enc_ring_get_rptr(struct amdgpu_ring *ring)
{
	struct amdgpu_device *adev = ring->adev;

	if (ring == &adev->vcn.ring_enc[0])
895
		return RREG32_SOC15(UVD, 0, mmUVD_RB_RPTR);
896
	else
897
		return RREG32_SOC15(UVD, 0, mmUVD_RB_RPTR2);
898 899 900 901 902 903 904 905 906 907 908 909 910 911
}

 /**
 * vcn_v1_0_enc_ring_get_wptr - get enc write pointer
 *
 * @ring: amdgpu_ring pointer
 *
 * Returns the current hardware enc write pointer
 */
static uint64_t vcn_v1_0_enc_ring_get_wptr(struct amdgpu_ring *ring)
{
	struct amdgpu_device *adev = ring->adev;

	if (ring == &adev->vcn.ring_enc[0])
912
		return RREG32_SOC15(UVD, 0, mmUVD_RB_WPTR);
913
	else
914
		return RREG32_SOC15(UVD, 0, mmUVD_RB_WPTR2);
915 916 917 918 919 920 921 922 923 924 925 926 927 928
}

 /**
 * vcn_v1_0_enc_ring_set_wptr - set enc write pointer
 *
 * @ring: amdgpu_ring pointer
 *
 * Commits the enc write pointer to the hardware
 */
static void vcn_v1_0_enc_ring_set_wptr(struct amdgpu_ring *ring)
{
	struct amdgpu_device *adev = ring->adev;

	if (ring == &adev->vcn.ring_enc[0])
929
		WREG32_SOC15(UVD, 0, mmUVD_RB_WPTR,
930 931
			lower_32_bits(ring->wptr));
	else
932
		WREG32_SOC15(UVD, 0, mmUVD_RB_WPTR2,
933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969
			lower_32_bits(ring->wptr));
}

/**
 * vcn_v1_0_enc_ring_emit_fence - emit an enc fence & trap command
 *
 * @ring: amdgpu_ring pointer
 * @fence: fence to emit
 *
 * Write enc a fence and a trap command to the ring.
 */
static void vcn_v1_0_enc_ring_emit_fence(struct amdgpu_ring *ring, u64 addr,
			u64 seq, unsigned flags)
{
	WARN_ON(flags & AMDGPU_FENCE_FLAG_64BIT);

	amdgpu_ring_write(ring, VCN_ENC_CMD_FENCE);
	amdgpu_ring_write(ring, addr);
	amdgpu_ring_write(ring, upper_32_bits(addr));
	amdgpu_ring_write(ring, seq);
	amdgpu_ring_write(ring, VCN_ENC_CMD_TRAP);
}

static void vcn_v1_0_enc_ring_insert_end(struct amdgpu_ring *ring)
{
	amdgpu_ring_write(ring, VCN_ENC_CMD_END);
}

/**
 * vcn_v1_0_enc_ring_emit_ib - enc execute indirect buffer
 *
 * @ring: amdgpu_ring pointer
 * @ib: indirect buffer to execute
 *
 * Write enc ring commands to execute the indirect buffer
 */
static void vcn_v1_0_enc_ring_emit_ib(struct amdgpu_ring *ring,
970
		struct amdgpu_ib *ib, unsigned int vmid, bool ctx_switch)
971 972
{
	amdgpu_ring_write(ring, VCN_ENC_CMD_IB);
973
	amdgpu_ring_write(ring, vmid);
974 975 976 977 978 979
	amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr));
	amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
	amdgpu_ring_write(ring, ib->length_dw);
}

static void vcn_v1_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
980 981
					    unsigned int vmid, unsigned pasid,
					    uint64_t pd_addr)
982 983 984 985
{
	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
	unsigned eng = ring->vm_inv_eng;

986
	pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
987

988
	/* wait for reg writes */
989 990
	amdgpu_ring_write(ring, VCN_ENC_CMD_REG_WAIT);
	amdgpu_ring_write(ring,
991
			  (hub->ctx0_ptb_addr_lo32 + vmid * 2) << 2);
992 993 994 995 996 997
	amdgpu_ring_write(ring, 0xffffffff);
	amdgpu_ring_write(ring, lower_32_bits(pd_addr));

	/* wait for flush */
	amdgpu_ring_write(ring, VCN_ENC_CMD_REG_WAIT);
	amdgpu_ring_write(ring, (hub->vm_inv_eng0_ack + eng) << 2);
998 999
	amdgpu_ring_write(ring, 1 << vmid);
	amdgpu_ring_write(ring, 1 << vmid);
1000 1001
}

1002 1003 1004 1005 1006 1007 1008 1009
static void vcn_v1_0_enc_ring_emit_wreg(struct amdgpu_ring *ring,
					uint32_t reg, uint32_t val)
{
	amdgpu_ring_write(ring, VCN_ENC_CMD_REG_WRITE);
	amdgpu_ring_write(ring,	reg << 2);
	amdgpu_ring_write(ring, val);
}

L
Leo Liu 已提交
1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023
static int vcn_v1_0_set_interrupt_state(struct amdgpu_device *adev,
					struct amdgpu_irq_src *source,
					unsigned type,
					enum amdgpu_interrupt_state state)
{
	return 0;
}

static int vcn_v1_0_process_interrupt(struct amdgpu_device *adev,
				      struct amdgpu_irq_src *source,
				      struct amdgpu_iv_entry *entry)
{
	DRM_DEBUG("IH: VCN TRAP\n");

L
Leo Liu 已提交
1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038
	switch (entry->src_id) {
	case 124:
		amdgpu_fence_process(&adev->vcn.ring_dec);
		break;
	case 119:
		amdgpu_fence_process(&adev->vcn.ring_enc[0]);
		break;
	case 120:
		amdgpu_fence_process(&adev->vcn.ring_enc[1]);
		break;
	default:
		DRM_ERROR("Unhandled interrupt: %d %d\n",
			  entry->src_id, entry->src_data[0]);
		break;
	}
L
Leo Liu 已提交
1039 1040 1041 1042

	return 0;
}

1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053
static void vcn_v1_0_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
{
	int i;
	struct amdgpu_device *adev = ring->adev;

	for (i = 0; i < count; i++)
		amdgpu_ring_write(ring, PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_NO_OP), 0));

}


1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072
static const struct amd_ip_funcs vcn_v1_0_ip_funcs = {
	.name = "vcn_v1_0",
	.early_init = vcn_v1_0_early_init,
	.late_init = NULL,
	.sw_init = vcn_v1_0_sw_init,
	.sw_fini = vcn_v1_0_sw_fini,
	.hw_init = vcn_v1_0_hw_init,
	.hw_fini = vcn_v1_0_hw_fini,
	.suspend = vcn_v1_0_suspend,
	.resume = vcn_v1_0_resume,
	.is_idle = NULL /* vcn_v1_0_is_idle */,
	.wait_for_idle = NULL /* vcn_v1_0_wait_for_idle */,
	.check_soft_reset = NULL /* vcn_v1_0_check_soft_reset */,
	.pre_soft_reset = NULL /* vcn_v1_0_pre_soft_reset */,
	.soft_reset = NULL /* vcn_v1_0_soft_reset */,
	.post_soft_reset = NULL /* vcn_v1_0_post_soft_reset */,
	.set_clockgating_state = vcn_v1_0_set_clockgating_state,
	.set_powergating_state = NULL /* vcn_v1_0_set_powergating_state */,
};
1073 1074 1075 1076

static const struct amdgpu_ring_funcs vcn_v1_0_dec_ring_vm_funcs = {
	.type = AMDGPU_RING_TYPE_VCN_DEC,
	.align_mask = 0xf,
1077
	.nop = PACKET0(0x81ff, 0),
1078
	.support_64bit_ptrs = false,
1079
	.vmhub = AMDGPU_MMHUB,
1080 1081 1082 1083
	.get_rptr = vcn_v1_0_dec_ring_get_rptr,
	.get_wptr = vcn_v1_0_dec_ring_get_wptr,
	.set_wptr = vcn_v1_0_dec_ring_set_wptr,
	.emit_frame_size =
1084
		6 + 6 + /* hdp invalidate / flush */
1085
		SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 + 16 + /* vcn_v1_0_dec_ring_emit_vm_flush */
1086
		14 + 14 + /* vcn_v1_0_dec_ring_emit_fence x2 vm fence */
1087
		6,
1088 1089 1090 1091
	.emit_ib_size = 8, /* vcn_v1_0_dec_ring_emit_ib */
	.emit_ib = vcn_v1_0_dec_ring_emit_ib,
	.emit_fence = vcn_v1_0_dec_ring_emit_fence,
	.emit_vm_flush = vcn_v1_0_dec_ring_emit_vm_flush,
1092
	.test_ring = amdgpu_vcn_dec_ring_test_ring,
1093
	.test_ib = amdgpu_vcn_dec_ring_test_ib,
1094
	.insert_nop = vcn_v1_0_ring_insert_nop,
1095
	.insert_start = vcn_v1_0_dec_ring_insert_start,
1096
	.insert_end = vcn_v1_0_dec_ring_insert_end,
1097 1098 1099
	.pad_ib = amdgpu_ring_generic_pad_ib,
	.begin_use = amdgpu_vcn_ring_begin_use,
	.end_use = amdgpu_vcn_ring_end_use,
1100
	.emit_wreg = vcn_v1_0_enc_ring_emit_wreg,
1101 1102
};

1103 1104 1105 1106 1107
static const struct amdgpu_ring_funcs vcn_v1_0_enc_ring_vm_funcs = {
	.type = AMDGPU_RING_TYPE_VCN_ENC,
	.align_mask = 0x3f,
	.nop = VCN_ENC_CMD_NO_OP,
	.support_64bit_ptrs = false,
1108
	.vmhub = AMDGPU_MMHUB,
1109 1110 1111 1112
	.get_rptr = vcn_v1_0_enc_ring_get_rptr,
	.get_wptr = vcn_v1_0_enc_ring_get_wptr,
	.set_wptr = vcn_v1_0_enc_ring_set_wptr,
	.emit_frame_size =
1113
		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 + 8 + /* vcn_v1_0_enc_ring_emit_vm_flush */
1114 1115 1116 1117 1118 1119
		5 + 5 + /* vcn_v1_0_enc_ring_emit_fence x2 vm fence */
		1, /* vcn_v1_0_enc_ring_insert_end */
	.emit_ib_size = 5, /* vcn_v1_0_enc_ring_emit_ib */
	.emit_ib = vcn_v1_0_enc_ring_emit_ib,
	.emit_fence = vcn_v1_0_enc_ring_emit_fence,
	.emit_vm_flush = vcn_v1_0_enc_ring_emit_vm_flush,
1120
	.test_ring = amdgpu_vcn_enc_ring_test_ring,
L
Leo Liu 已提交
1121
	.test_ib = amdgpu_vcn_enc_ring_test_ib,
1122 1123 1124 1125 1126 1127 1128
	.insert_nop = amdgpu_ring_insert_nop,
	.insert_end = vcn_v1_0_enc_ring_insert_end,
	.pad_ib = amdgpu_ring_generic_pad_ib,
	.begin_use = amdgpu_vcn_ring_begin_use,
	.end_use = amdgpu_vcn_ring_end_use,
};

1129 1130 1131 1132 1133
static void vcn_v1_0_set_dec_ring_funcs(struct amdgpu_device *adev)
{
	adev->vcn.ring_dec.funcs = &vcn_v1_0_dec_ring_vm_funcs;
	DRM_INFO("VCN decode is enabled in VM mode\n");
}
L
Leo Liu 已提交
1134

1135 1136 1137 1138 1139 1140 1141 1142 1143 1144
static void vcn_v1_0_set_enc_ring_funcs(struct amdgpu_device *adev)
{
	int i;

	for (i = 0; i < adev->vcn.num_enc_rings; ++i)
		adev->vcn.ring_enc[i].funcs = &vcn_v1_0_enc_ring_vm_funcs;

	DRM_INFO("VCN encode is enabled in VM mode\n");
}

L
Leo Liu 已提交
1145 1146 1147 1148 1149 1150 1151
static const struct amdgpu_irq_src_funcs vcn_v1_0_irq_funcs = {
	.set = vcn_v1_0_set_interrupt_state,
	.process = vcn_v1_0_process_interrupt,
};

static void vcn_v1_0_set_irq_funcs(struct amdgpu_device *adev)
{
1152
	adev->vcn.irq.num_types = adev->vcn.num_enc_rings + 1;
L
Leo Liu 已提交
1153 1154
	adev->vcn.irq.funcs = &vcn_v1_0_irq_funcs;
}
L
Leo Liu 已提交
1155 1156 1157 1158 1159 1160 1161 1162 1163

const struct amdgpu_ip_block_version vcn_v1_0_ip_block =
{
		.type = AMD_IP_BLOCK_TYPE_VCN,
		.major = 1,
		.minor = 0,
		.rev = 0,
		.funcs = &vcn_v1_0_ip_funcs,
};