vc4_plane.c 24.2 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
/*
 * Copyright (C) 2015 Broadcom
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as
 * published by the Free Software Foundation.
 */

/**
 * DOC: VC4 plane module
 *
 * Each DRM plane is a layer of pixels being scanned out by the HVS.
 *
 * At atomic modeset check time, we compute the HVS display element
 * state that would be necessary for displaying the plane (giving us a
 * chance to figure out if a plane configuration is invalid), then at
 * atomic flush time the CRTC will ask us to write our element state
 * into the region of the HVS that it has allocated for us.
 */

#include "vc4_drv.h"
#include "vc4_regs.h"
#include "drm_atomic_helper.h"
#include "drm_fb_cma_helper.h"
#include "drm_plane_helper.h"

27 28 29 30 31 32
enum vc4_scaling_mode {
	VC4_SCALING_NONE,
	VC4_SCALING_TPZ,
	VC4_SCALING_PPF,
};

33 34
struct vc4_plane_state {
	struct drm_plane_state base;
35 36 37
	/* System memory copy of the display list for this element, computed
	 * at atomic_check time.
	 */
38
	u32 *dlist;
39
	u32 dlist_size; /* Number of dwords allocated for the display list */
40
	u32 dlist_count; /* Number of used dwords in the display list. */
41

42 43 44 45 46 47
	/* Offset in the dlist to various words, for pageflip or
	 * cursor updates.
	 */
	u32 pos0_offset;
	u32 pos2_offset;
	u32 ptr0_offset;
48 49

	/* Offset where the plane's dlist was last stored in the
50 51
	 * hardware at vc4_crtc_atomic_flush() time.
	 */
52
	u32 __iomem *hw_dlist;
53 54 55

	/* Clipped coordinates of the plane on the display. */
	int crtc_x, crtc_y, crtc_w, crtc_h;
56
	/* Clipped area being scanned from in the FB. */
57
	u32 src_x, src_y;
58

59 60 61 62
	u32 src_w[2], src_h[2];

	/* Scaling selection for the RGB/Y plane and the Cb/Cr planes. */
	enum vc4_scaling_mode x_scaling[2], y_scaling[2];
63
	bool is_unity;
64
	bool is_yuv;
65 66 67 68

	/* Offset to start scanning out from the start of the plane's
	 * BO.
	 */
69
	u32 offsets[3];
70 71 72

	/* Our allocation in LBM for temporary storage during scaling. */
	struct drm_mm_node lbm;
73 74 75 76 77 78 79 80 81 82 83 84 85
};

static inline struct vc4_plane_state *
to_vc4_plane_state(struct drm_plane_state *state)
{
	return (struct vc4_plane_state *)state;
}

static const struct hvs_format {
	u32 drm; /* DRM_FORMAT_* */
	u32 hvs; /* HVS_FORMAT_* */
	u32 pixel_order;
	bool has_alpha;
86
	bool flip_cbcr;
87 88 89 90 91 92 93 94 95
} hvs_formats[] = {
	{
		.drm = DRM_FORMAT_XRGB8888, .hvs = HVS_PIXEL_FORMAT_RGBA8888,
		.pixel_order = HVS_PIXEL_ORDER_ABGR, .has_alpha = false,
	},
	{
		.drm = DRM_FORMAT_ARGB8888, .hvs = HVS_PIXEL_FORMAT_RGBA8888,
		.pixel_order = HVS_PIXEL_ORDER_ABGR, .has_alpha = true,
	},
96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111
	{
		.drm = DRM_FORMAT_RGB565, .hvs = HVS_PIXEL_FORMAT_RGB565,
		.pixel_order = HVS_PIXEL_ORDER_XRGB, .has_alpha = false,
	},
	{
		.drm = DRM_FORMAT_BGR565, .hvs = HVS_PIXEL_FORMAT_RGB565,
		.pixel_order = HVS_PIXEL_ORDER_XBGR, .has_alpha = false,
	},
	{
		.drm = DRM_FORMAT_ARGB1555, .hvs = HVS_PIXEL_FORMAT_RGBA5551,
		.pixel_order = HVS_PIXEL_ORDER_ABGR, .has_alpha = true,
	},
	{
		.drm = DRM_FORMAT_XRGB1555, .hvs = HVS_PIXEL_FORMAT_RGBA5551,
		.pixel_order = HVS_PIXEL_ORDER_ABGR, .has_alpha = false,
	},
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
	{
		.drm = DRM_FORMAT_YUV422,
		.hvs = HVS_PIXEL_FORMAT_YCBCR_YUV422_3PLANE,
	},
	{
		.drm = DRM_FORMAT_YVU422,
		.hvs = HVS_PIXEL_FORMAT_YCBCR_YUV422_3PLANE,
		.flip_cbcr = true,
	},
	{
		.drm = DRM_FORMAT_YUV420,
		.hvs = HVS_PIXEL_FORMAT_YCBCR_YUV420_3PLANE,
	},
	{
		.drm = DRM_FORMAT_YVU420,
		.hvs = HVS_PIXEL_FORMAT_YCBCR_YUV420_3PLANE,
		.flip_cbcr = true,
	},
	{
		.drm = DRM_FORMAT_NV12,
		.hvs = HVS_PIXEL_FORMAT_YCBCR_YUV420_2PLANE,
	},
	{
		.drm = DRM_FORMAT_NV16,
		.hvs = HVS_PIXEL_FORMAT_YCBCR_YUV422_2PLANE,
	},
138 139 140 141 142 143 144 145 146 147 148 149 150 151
};

static const struct hvs_format *vc4_get_hvs_format(u32 drm_format)
{
	unsigned i;

	for (i = 0; i < ARRAY_SIZE(hvs_formats); i++) {
		if (hvs_formats[i].drm == drm_format)
			return &hvs_formats[i];
	}

	return NULL;
}

152 153 154 155 156 157 158 159 160 161
static enum vc4_scaling_mode vc4_get_scaling_mode(u32 src, u32 dst)
{
	if (dst > src)
		return VC4_SCALING_PPF;
	else if (dst < src)
		return VC4_SCALING_TPZ;
	else
		return VC4_SCALING_NONE;
}

162 163 164 165 166
static bool plane_enabled(struct drm_plane_state *state)
{
	return state->fb && state->crtc;
}

167
static struct drm_plane_state *vc4_plane_duplicate_state(struct drm_plane *plane)
168 169 170 171 172 173 174 175 176 177
{
	struct vc4_plane_state *vc4_state;

	if (WARN_ON(!plane->state))
		return NULL;

	vc4_state = kmemdup(plane->state, sizeof(*vc4_state), GFP_KERNEL);
	if (!vc4_state)
		return NULL;

178 179
	memset(&vc4_state->lbm, 0, sizeof(vc4_state->lbm));

180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195
	__drm_atomic_helper_plane_duplicate_state(plane, &vc4_state->base);

	if (vc4_state->dlist) {
		vc4_state->dlist = kmemdup(vc4_state->dlist,
					   vc4_state->dlist_count * 4,
					   GFP_KERNEL);
		if (!vc4_state->dlist) {
			kfree(vc4_state);
			return NULL;
		}
		vc4_state->dlist_size = vc4_state->dlist_count;
	}

	return &vc4_state->base;
}

196 197
static void vc4_plane_destroy_state(struct drm_plane *plane,
				    struct drm_plane_state *state)
198
{
199
	struct vc4_dev *vc4 = to_vc4_dev(plane->dev);
200 201
	struct vc4_plane_state *vc4_state = to_vc4_plane_state(state);

202 203 204 205 206 207 208 209
	if (vc4_state->lbm.allocated) {
		unsigned long irqflags;

		spin_lock_irqsave(&vc4->hvs->mm_lock, irqflags);
		drm_mm_remove_node(&vc4_state->lbm);
		spin_unlock_irqrestore(&vc4->hvs->mm_lock, irqflags);
	}

210 211 212 213 214 215
	kfree(vc4_state->dlist);
	__drm_atomic_helper_plane_destroy_state(plane, &vc4_state->base);
	kfree(state);
}

/* Called during init to allocate the plane's atomic state. */
216
static void vc4_plane_reset(struct drm_plane *plane)
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
{
	struct vc4_plane_state *vc4_state;

	WARN_ON(plane->state);

	vc4_state = kzalloc(sizeof(*vc4_state), GFP_KERNEL);
	if (!vc4_state)
		return;

	plane->state = &vc4_state->base;
	vc4_state->base.plane = plane;
}

static void vc4_dlist_write(struct vc4_plane_state *vc4_state, u32 val)
{
	if (vc4_state->dlist_count == vc4_state->dlist_size) {
		u32 new_size = max(4u, vc4_state->dlist_count * 2);
		u32 *new_dlist = kmalloc(new_size * 4, GFP_KERNEL);

		if (!new_dlist)
			return;
		memcpy(new_dlist, vc4_state->dlist, vc4_state->dlist_count * 4);

		kfree(vc4_state->dlist);
		vc4_state->dlist = new_dlist;
		vc4_state->dlist_size = new_size;
	}

	vc4_state->dlist[vc4_state->dlist_count++] = val;
}

248 249 250 251 252
/* Returns the scl0/scl1 field based on whether the dimensions need to
 * be up/down/non-scaled.
 *
 * This is a replication of a table from the spec.
 */
253
static u32 vc4_get_scl_field(struct drm_plane_state *state, int plane)
254 255 256
{
	struct vc4_plane_state *vc4_state = to_vc4_plane_state(state);

257
	switch (vc4_state->x_scaling[plane] << 2 | vc4_state->y_scaling[plane]) {
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
	case VC4_SCALING_PPF << 2 | VC4_SCALING_PPF:
		return SCALER_CTL0_SCL_H_PPF_V_PPF;
	case VC4_SCALING_TPZ << 2 | VC4_SCALING_PPF:
		return SCALER_CTL0_SCL_H_TPZ_V_PPF;
	case VC4_SCALING_PPF << 2 | VC4_SCALING_TPZ:
		return SCALER_CTL0_SCL_H_PPF_V_TPZ;
	case VC4_SCALING_TPZ << 2 | VC4_SCALING_TPZ:
		return SCALER_CTL0_SCL_H_TPZ_V_TPZ;
	case VC4_SCALING_PPF << 2 | VC4_SCALING_NONE:
		return SCALER_CTL0_SCL_H_PPF_V_NONE;
	case VC4_SCALING_NONE << 2 | VC4_SCALING_PPF:
		return SCALER_CTL0_SCL_H_NONE_V_PPF;
	case VC4_SCALING_NONE << 2 | VC4_SCALING_TPZ:
		return SCALER_CTL0_SCL_H_NONE_V_TPZ;
	case VC4_SCALING_TPZ << 2 | VC4_SCALING_NONE:
		return SCALER_CTL0_SCL_H_TPZ_V_NONE;
	default:
	case VC4_SCALING_NONE << 2 | VC4_SCALING_NONE:
		/* The unity case is independently handled by
		 * SCALER_CTL0_UNITY.
		 */
		return 0;
	}
}

283
static int vc4_plane_setup_clipping_and_scaling(struct drm_plane_state *state)
284
{
285
	struct drm_plane *plane = state->plane;
286 287
	struct vc4_plane_state *vc4_state = to_vc4_plane_state(state);
	struct drm_framebuffer *fb = state->fb;
288
	struct drm_gem_cma_object *bo = drm_fb_cma_get_gem_obj(fb, 0);
289
	u32 subpixel_src_mask = (1 << 16) - 1;
290 291 292 293 294
	u32 format = fb->pixel_format;
	int num_planes = drm_format_num_planes(format);
	u32 h_subsample = 1;
	u32 v_subsample = 1;
	int i;
295

296 297
	for (i = 0; i < num_planes; i++)
		vc4_state->offsets[i] = bo->paddr + fb->offsets[i];
298

299 300 301 302 303
	/* We don't support subpixel source positioning for scaling. */
	if ((state->src_x & subpixel_src_mask) ||
	    (state->src_y & subpixel_src_mask) ||
	    (state->src_w & subpixel_src_mask) ||
	    (state->src_h & subpixel_src_mask)) {
304 305 306
		return -EINVAL;
	}

307 308
	vc4_state->src_x = state->src_x >> 16;
	vc4_state->src_y = state->src_y >> 16;
309 310
	vc4_state->src_w[0] = state->src_w >> 16;
	vc4_state->src_h[0] = state->src_h >> 16;
311 312 313 314 315 316

	vc4_state->crtc_x = state->crtc_x;
	vc4_state->crtc_y = state->crtc_y;
	vc4_state->crtc_w = state->crtc_w;
	vc4_state->crtc_h = state->crtc_h;

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
	vc4_state->x_scaling[0] = vc4_get_scaling_mode(vc4_state->src_w[0],
						       vc4_state->crtc_w);
	vc4_state->y_scaling[0] = vc4_get_scaling_mode(vc4_state->src_h[0],
						       vc4_state->crtc_h);

	if (num_planes > 1) {
		vc4_state->is_yuv = true;

		h_subsample = drm_format_horz_chroma_subsampling(format);
		v_subsample = drm_format_vert_chroma_subsampling(format);
		vc4_state->src_w[1] = vc4_state->src_w[0] / h_subsample;
		vc4_state->src_h[1] = vc4_state->src_h[0] / v_subsample;

		vc4_state->x_scaling[1] =
			vc4_get_scaling_mode(vc4_state->src_w[1],
					     vc4_state->crtc_w);
		vc4_state->y_scaling[1] =
			vc4_get_scaling_mode(vc4_state->src_h[1],
					     vc4_state->crtc_h);

		/* YUV conversion requires that scaling be enabled,
		 * even on a plane that's otherwise 1:1.  Choose TPZ
		 * for simplicity.
		 */
		if (vc4_state->x_scaling[0] == VC4_SCALING_NONE)
			vc4_state->x_scaling[0] = VC4_SCALING_TPZ;
		if (vc4_state->y_scaling[0] == VC4_SCALING_NONE)
			vc4_state->y_scaling[0] = VC4_SCALING_TPZ;
	}

	vc4_state->is_unity = (vc4_state->x_scaling[0] == VC4_SCALING_NONE &&
			       vc4_state->y_scaling[0] == VC4_SCALING_NONE &&
			       vc4_state->x_scaling[1] == VC4_SCALING_NONE &&
			       vc4_state->y_scaling[1] == VC4_SCALING_NONE);
351 352 353 354 355 356 357 358 359 360 361

	/* No configuring scaling on the cursor plane, since it gets
	   non-vblank-synced updates, and scaling requires requires
	   LBM changes which have to be vblank-synced.
	 */
	if (plane->type == DRM_PLANE_TYPE_CURSOR && !vc4_state->is_unity)
		return -EINVAL;

	/* Clamp the on-screen start x/y to 0.  The hardware doesn't
	 * support negative y, and negative x wastes bandwidth.
	 */
362
	if (vc4_state->crtc_x < 0) {
363 364 365 366 367 368 369 370 371
		for (i = 0; i < num_planes; i++) {
			u32 cpp = drm_format_plane_cpp(fb->pixel_format, i);
			u32 subs = ((i == 0) ? 1 : h_subsample);

			vc4_state->offsets[i] += (cpp *
						  (-vc4_state->crtc_x) / subs);
		}
		vc4_state->src_w[0] += vc4_state->crtc_x;
		vc4_state->src_w[1] += vc4_state->crtc_x / h_subsample;
372
		vc4_state->crtc_x = 0;
373 374
	}

375
	if (vc4_state->crtc_y < 0) {
376 377 378 379 380 381 382 383
		for (i = 0; i < num_planes; i++) {
			u32 subs = ((i == 0) ? 1 : v_subsample);

			vc4_state->offsets[i] += (fb->pitches[i] *
						  (-vc4_state->crtc_y) / subs);
		}
		vc4_state->src_h[0] += vc4_state->crtc_y;
		vc4_state->src_h[1] += vc4_state->crtc_y / v_subsample;
384
		vc4_state->crtc_y = 0;
385 386
	}

387 388 389
	return 0;
}

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
static void vc4_write_tpz(struct vc4_plane_state *vc4_state, u32 src, u32 dst)
{
	u32 scale, recip;

	scale = (1 << 16) * src / dst;

	/* The specs note that while the reciprocal would be defined
	 * as (1<<32)/scale, ~0 is close enough.
	 */
	recip = ~0 / scale;

	vc4_dlist_write(vc4_state,
			VC4_SET_FIELD(scale, SCALER_TPZ0_SCALE) |
			VC4_SET_FIELD(0, SCALER_TPZ0_IPHASE));
	vc4_dlist_write(vc4_state,
			VC4_SET_FIELD(recip, SCALER_TPZ1_RECIP));
}

static void vc4_write_ppf(struct vc4_plane_state *vc4_state, u32 src, u32 dst)
{
	u32 scale = (1 << 16) * src / dst;

	vc4_dlist_write(vc4_state,
			SCALER_PPF_AGC |
			VC4_SET_FIELD(scale, SCALER_PPF_SCALE) |
			VC4_SET_FIELD(0, SCALER_PPF_IPHASE));
}

static u32 vc4_lbm_size(struct drm_plane_state *state)
{
	struct vc4_plane_state *vc4_state = to_vc4_plane_state(state);
	/* This is the worst case number.  One of the two sizes will
	 * be used depending on the scaling configuration.
	 */
424
	u32 pix_per_line = max(vc4_state->src_w[0], (u32)vc4_state->crtc_w);
425 426
	u32 lbm;

427 428 429 430 431 432 433 434 435 436 437 438 439 440
	if (!vc4_state->is_yuv) {
		if (vc4_state->is_unity)
			return 0;
		else if (vc4_state->y_scaling[0] == VC4_SCALING_TPZ)
			lbm = pix_per_line * 8;
		else {
			/* In special cases, this multiplier might be 12. */
			lbm = pix_per_line * 16;
		}
	} else {
		/* There are cases for this going down to a multiplier
		 * of 2, but according to the firmware source, the
		 * table in the docs is somewhat wrong.
		 */
441 442 443 444 445 446 447 448
		lbm = pix_per_line * 16;
	}

	lbm = roundup(lbm, 32);

	return lbm;
}

449 450
static void vc4_write_scaling_parameters(struct drm_plane_state *state,
					 int channel)
451 452 453 454
{
	struct vc4_plane_state *vc4_state = to_vc4_plane_state(state);

	/* Ch0 H-PPF Word 0: Scaling Parameters */
455
	if (vc4_state->x_scaling[channel] == VC4_SCALING_PPF) {
456
		vc4_write_ppf(vc4_state,
457
			      vc4_state->src_w[channel], vc4_state->crtc_w);
458 459 460
	}

	/* Ch0 V-PPF Words 0-1: Scaling Parameters, Context */
461
	if (vc4_state->y_scaling[channel] == VC4_SCALING_PPF) {
462
		vc4_write_ppf(vc4_state,
463
			      vc4_state->src_h[channel], vc4_state->crtc_h);
464 465 466 467
		vc4_dlist_write(vc4_state, 0xc0c0c0c0);
	}

	/* Ch0 H-TPZ Words 0-1: Scaling Parameters, Recip */
468
	if (vc4_state->x_scaling[channel] == VC4_SCALING_TPZ) {
469
		vc4_write_tpz(vc4_state,
470
			      vc4_state->src_w[channel], vc4_state->crtc_w);
471 472 473
	}

	/* Ch0 V-TPZ Words 0-2: Scaling Parameters, Recip, Context */
474
	if (vc4_state->y_scaling[channel] == VC4_SCALING_TPZ) {
475
		vc4_write_tpz(vc4_state,
476
			      vc4_state->src_h[channel], vc4_state->crtc_h);
477 478 479
		vc4_dlist_write(vc4_state, 0xc0c0c0c0);
	}
}
480 481 482 483 484 485 486

/* Writes out a full display list for an active plane to the plane's
 * private dlist state.
 */
static int vc4_plane_mode_set(struct drm_plane *plane,
			      struct drm_plane_state *state)
{
487
	struct vc4_dev *vc4 = to_vc4_dev(plane->dev);
488 489 490 491
	struct vc4_plane_state *vc4_state = to_vc4_plane_state(state);
	struct drm_framebuffer *fb = state->fb;
	u32 ctl0_offset = vc4_state->dlist_count;
	const struct hvs_format *format = vc4_get_hvs_format(fb->pixel_format);
492 493
	int num_planes = drm_format_num_planes(format->drm);
	u32 scl0, scl1;
494 495
	u32 lbm_size;
	unsigned long irqflags;
496
	int ret, i;
497 498 499 500 501

	ret = vc4_plane_setup_clipping_and_scaling(state);
	if (ret)
		return ret;

502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520
	/* Allocate the LBM memory that the HVS will use for temporary
	 * storage due to our scaling/format conversion.
	 */
	lbm_size = vc4_lbm_size(state);
	if (lbm_size) {
		if (!vc4_state->lbm.allocated) {
			spin_lock_irqsave(&vc4->hvs->mm_lock, irqflags);
			ret = drm_mm_insert_node(&vc4->hvs->lbm_mm,
						 &vc4_state->lbm,
						 lbm_size, 32, 0);
			spin_unlock_irqrestore(&vc4->hvs->mm_lock, irqflags);
		} else {
			WARN_ON_ONCE(lbm_size != vc4_state->lbm.size);
		}
	}

	if (ret)
		return ret;

521 522 523 524 525 526 527 528 529 530 531 532 533
	/* SCL1 is used for Cb/Cr scaling of planar formats.  For RGB
	 * and 4:4:4, scl1 should be set to scl0 so both channels of
	 * the scaler do the same thing.  For YUV, the Y plane needs
	 * to be put in channel 1 and Cb/Cr in channel 0, so we swap
	 * the scl fields here.
	 */
	if (num_planes == 1) {
		scl0 = vc4_get_scl_field(state, 1);
		scl1 = scl0;
	} else {
		scl0 = vc4_get_scl_field(state, 1);
		scl1 = vc4_get_scl_field(state, 0);
	}
534 535

	/* Control word */
536 537 538 539
	vc4_dlist_write(vc4_state,
			SCALER_CTL0_VALID |
			(format->pixel_order << SCALER_CTL0_ORDER_SHIFT) |
			(format->hvs << SCALER_CTL0_PIXEL_FORMAT_SHIFT) |
540
			(vc4_state->is_unity ? SCALER_CTL0_UNITY : 0) |
541 542
			VC4_SET_FIELD(scl0, SCALER_CTL0_SCL0) |
			VC4_SET_FIELD(scl1, SCALER_CTL0_SCL1));
543 544

	/* Position Word 0: Image Positions and Alpha Value */
545
	vc4_state->pos0_offset = vc4_state->dlist_count;
546 547
	vc4_dlist_write(vc4_state,
			VC4_SET_FIELD(0xff, SCALER_POS0_FIXED_ALPHA) |
548 549
			VC4_SET_FIELD(vc4_state->crtc_x, SCALER_POS0_START_X) |
			VC4_SET_FIELD(vc4_state->crtc_y, SCALER_POS0_START_Y));
550

551 552 553 554 555 556 557 558
	/* Position Word 1: Scaled Image Dimensions. */
	if (!vc4_state->is_unity) {
		vc4_dlist_write(vc4_state,
				VC4_SET_FIELD(vc4_state->crtc_w,
					      SCALER_POS1_SCL_WIDTH) |
				VC4_SET_FIELD(vc4_state->crtc_h,
					      SCALER_POS1_SCL_HEIGHT));
	}
559 560

	/* Position Word 2: Source Image Size, Alpha Mode */
561
	vc4_state->pos2_offset = vc4_state->dlist_count;
562 563 564 565 566
	vc4_dlist_write(vc4_state,
			VC4_SET_FIELD(format->has_alpha ?
				      SCALER_POS2_ALPHA_MODE_PIPELINE :
				      SCALER_POS2_ALPHA_MODE_FIXED,
				      SCALER_POS2_ALPHA_MODE) |
567 568
			VC4_SET_FIELD(vc4_state->src_w[0], SCALER_POS2_WIDTH) |
			VC4_SET_FIELD(vc4_state->src_h[0], SCALER_POS2_HEIGHT));
569 570 571 572

	/* Position Word 3: Context.  Written by the HVS. */
	vc4_dlist_write(vc4_state, 0xc0c0c0c0);

573 574 575 576 577

	/* Pointer Word 0/1/2: RGB / Y / Cb / Cr Pointers
	 *
	 * The pointers may be any byte address.
	 */
578
	vc4_state->ptr0_offset = vc4_state->dlist_count;
579 580 581 582 583 584 585 586 587
	if (!format->flip_cbcr) {
		for (i = 0; i < num_planes; i++)
			vc4_dlist_write(vc4_state, vc4_state->offsets[i]);
	} else {
		WARN_ON_ONCE(num_planes != 3);
		vc4_dlist_write(vc4_state, vc4_state->offsets[0]);
		vc4_dlist_write(vc4_state, vc4_state->offsets[2]);
		vc4_dlist_write(vc4_state, vc4_state->offsets[1]);
	}
588

589 590 591
	/* Pointer Context Word 0/1/2: Written by the HVS */
	for (i = 0; i < num_planes; i++)
		vc4_dlist_write(vc4_state, 0xc0c0c0c0);
592

593 594 595 596 597 598 599 600 601 602 603 604
	/* Pitch word 0/1/2 */
	for (i = 0; i < num_planes; i++) {
		vc4_dlist_write(vc4_state,
				VC4_SET_FIELD(fb->pitches[i], SCALER_SRC_PITCH));
	}

	/* Colorspace conversion words */
	if (vc4_state->is_yuv) {
		vc4_dlist_write(vc4_state, SCALER_CSC0_ITR_R_601_5);
		vc4_dlist_write(vc4_state, SCALER_CSC1_ITR_R_601_5);
		vc4_dlist_write(vc4_state, SCALER_CSC2_ITR_R_601_5);
	}
605

606 607
	if (!vc4_state->is_unity) {
		/* LBM Base Address. */
608 609
		if (vc4_state->y_scaling[0] != VC4_SCALING_NONE ||
		    vc4_state->y_scaling[1] != VC4_SCALING_NONE) {
610
			vc4_dlist_write(vc4_state, vc4_state->lbm.start);
611
		}
612

613 614 615 616 617 618 619 620
		if (num_planes > 1) {
			/* Emit Cb/Cr as channel 0 and Y as channel
			 * 1. This matches how we set up scl0/scl1
			 * above.
			 */
			vc4_write_scaling_parameters(state, 1);
		}
		vc4_write_scaling_parameters(state, 0);
621 622 623 624

		/* If any PPF setup was done, then all the kernel
		 * pointers get uploaded.
		 */
625 626 627 628
		if (vc4_state->x_scaling[0] == VC4_SCALING_PPF ||
		    vc4_state->y_scaling[0] == VC4_SCALING_PPF ||
		    vc4_state->x_scaling[1] == VC4_SCALING_PPF ||
		    vc4_state->y_scaling[1] == VC4_SCALING_PPF) {
629 630 631 632 633 634 635 636 637 638 639 640 641 642
			u32 kernel = VC4_SET_FIELD(vc4->hvs->mitchell_netravali_filter.start,
						   SCALER_PPF_KERNEL_OFFSET);

			/* HPPF plane 0 */
			vc4_dlist_write(vc4_state, kernel);
			/* VPPF plane 0 */
			vc4_dlist_write(vc4_state, kernel);
			/* HPPF plane 1 */
			vc4_dlist_write(vc4_state, kernel);
			/* VPPF plane 1 */
			vc4_dlist_write(vc4_state, kernel);
		}
	}

643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683
	vc4_state->dlist[ctl0_offset] |=
		VC4_SET_FIELD(vc4_state->dlist_count, SCALER_CTL0_SIZE);

	return 0;
}

/* If a modeset involves changing the setup of a plane, the atomic
 * infrastructure will call this to validate a proposed plane setup.
 * However, if a plane isn't getting updated, this (and the
 * corresponding vc4_plane_atomic_update) won't get called.  Thus, we
 * compute the dlist here and have all active plane dlists get updated
 * in the CRTC's flush.
 */
static int vc4_plane_atomic_check(struct drm_plane *plane,
				  struct drm_plane_state *state)
{
	struct vc4_plane_state *vc4_state = to_vc4_plane_state(state);

	vc4_state->dlist_count = 0;

	if (plane_enabled(state))
		return vc4_plane_mode_set(plane, state);
	else
		return 0;
}

static void vc4_plane_atomic_update(struct drm_plane *plane,
				    struct drm_plane_state *old_state)
{
	/* No contents here.  Since we don't know where in the CRTC's
	 * dlist we should be stored, our dlist is uploaded to the
	 * hardware with vc4_plane_write_dlist() at CRTC atomic_flush
	 * time.
	 */
}

u32 vc4_plane_write_dlist(struct drm_plane *plane, u32 __iomem *dlist)
{
	struct vc4_plane_state *vc4_state = to_vc4_plane_state(plane->state);
	int i;

684 685
	vc4_state->hw_dlist = dlist;

686 687 688 689 690 691 692 693 694 695 696 697 698 699
	/* Can't memcpy_toio() because it needs to be 32-bit writes. */
	for (i = 0; i < vc4_state->dlist_count; i++)
		writel(vc4_state->dlist[i], &dlist[i]);

	return vc4_state->dlist_count;
}

u32 vc4_plane_dlist_size(struct drm_plane_state *state)
{
	struct vc4_plane_state *vc4_state = to_vc4_plane_state(state);

	return vc4_state->dlist_count;
}

700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718
/* Updates the plane to immediately (well, once the FIFO needs
 * refilling) scan out from at a new framebuffer.
 */
void vc4_plane_async_set_fb(struct drm_plane *plane, struct drm_framebuffer *fb)
{
	struct vc4_plane_state *vc4_state = to_vc4_plane_state(plane->state);
	struct drm_gem_cma_object *bo = drm_fb_cma_get_gem_obj(fb, 0);
	uint32_t addr;

	/* We're skipping the address adjustment for negative origin,
	 * because this is only called on the primary plane.
	 */
	WARN_ON_ONCE(plane->state->crtc_x < 0 || plane->state->crtc_y < 0);
	addr = bo->paddr + fb->offsets[0];

	/* Write the new address into the hardware immediately.  The
	 * scanout will start from this address as soon as the FIFO
	 * needs to refill with pixels.
	 */
719
	writel(addr, &vc4_state->hw_dlist[vc4_state->ptr0_offset]);
720 721 722 723 724

	/* Also update the CPU-side dlist copy, so that any later
	 * atomic updates that don't do a new modeset on our plane
	 * also use our updated address.
	 */
725
	vc4_state->dlist[vc4_state->ptr0_offset] = addr;
726 727
}

728 729 730 731 732 733 734 735 736 737 738 739 740
static const struct drm_plane_helper_funcs vc4_plane_helper_funcs = {
	.prepare_fb = NULL,
	.cleanup_fb = NULL,
	.atomic_check = vc4_plane_atomic_check,
	.atomic_update = vc4_plane_atomic_update,
};

static void vc4_plane_destroy(struct drm_plane *plane)
{
	drm_plane_helper_disable(plane);
	drm_plane_cleanup(plane);
}

741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815
/* Implements immediate (non-vblank-synced) updates of the cursor
 * position, or falls back to the atomic helper otherwise.
 */
static int
vc4_update_plane(struct drm_plane *plane,
		 struct drm_crtc *crtc,
		 struct drm_framebuffer *fb,
		 int crtc_x, int crtc_y,
		 unsigned int crtc_w, unsigned int crtc_h,
		 uint32_t src_x, uint32_t src_y,
		 uint32_t src_w, uint32_t src_h)
{
	struct drm_plane_state *plane_state;
	struct vc4_plane_state *vc4_state;

	if (plane != crtc->cursor)
		goto out;

	plane_state = plane->state;
	vc4_state = to_vc4_plane_state(plane_state);

	if (!plane_state)
		goto out;

	/* If we're changing the cursor contents, do that in the
	 * normal vblank-synced atomic path.
	 */
	if (fb != plane_state->fb)
		goto out;

	/* No configuring new scaling in the fast path. */
	if (crtc_w != plane_state->crtc_w ||
	    crtc_h != plane_state->crtc_h ||
	    src_w != plane_state->src_w ||
	    src_h != plane_state->src_h) {
		goto out;
	}

	/* Set the cursor's position on the screen.  This is the
	 * expected change from the drm_mode_cursor_universal()
	 * helper.
	 */
	plane_state->crtc_x = crtc_x;
	plane_state->crtc_y = crtc_y;

	/* Allow changing the start position within the cursor BO, if
	 * that matters.
	 */
	plane_state->src_x = src_x;
	plane_state->src_y = src_y;

	/* Update the display list based on the new crtc_x/y. */
	vc4_plane_atomic_check(plane, plane_state);

	/* Note that we can't just call vc4_plane_write_dlist()
	 * because that would smash the context data that the HVS is
	 * currently using.
	 */
	writel(vc4_state->dlist[vc4_state->pos0_offset],
	       &vc4_state->hw_dlist[vc4_state->pos0_offset]);
	writel(vc4_state->dlist[vc4_state->pos2_offset],
	       &vc4_state->hw_dlist[vc4_state->pos2_offset]);
	writel(vc4_state->dlist[vc4_state->ptr0_offset],
	       &vc4_state->hw_dlist[vc4_state->ptr0_offset]);

	return 0;

out:
	return drm_atomic_helper_update_plane(plane, crtc, fb,
					      crtc_x, crtc_y,
					      crtc_w, crtc_h,
					      src_x, src_y,
					      src_w, src_h);
}

816
static const struct drm_plane_funcs vc4_plane_funcs = {
817
	.update_plane = vc4_update_plane,
818 819 820 821 822 823 824 825 826 827 828 829 830 831
	.disable_plane = drm_atomic_helper_disable_plane,
	.destroy = vc4_plane_destroy,
	.set_property = NULL,
	.reset = vc4_plane_reset,
	.atomic_duplicate_state = vc4_plane_duplicate_state,
	.atomic_destroy_state = vc4_plane_destroy_state,
};

struct drm_plane *vc4_plane_init(struct drm_device *dev,
				 enum drm_plane_type type)
{
	struct drm_plane *plane = NULL;
	struct vc4_plane *vc4_plane;
	u32 formats[ARRAY_SIZE(hvs_formats)];
832
	u32 num_formats = 0;
833 834 835 836 837 838 839 840 841 842
	int ret = 0;
	unsigned i;

	vc4_plane = devm_kzalloc(dev->dev, sizeof(*vc4_plane),
				 GFP_KERNEL);
	if (!vc4_plane) {
		ret = -ENOMEM;
		goto fail;
	}

843 844 845 846 847 848 849 850 851 852
	for (i = 0; i < ARRAY_SIZE(hvs_formats); i++) {
		/* Don't allow YUV in cursor planes, since that means
		 * tuning on the scaler, which we don't allow for the
		 * cursor.
		 */
		if (type != DRM_PLANE_TYPE_CURSOR ||
		    hvs_formats[i].hvs < HVS_PIXEL_FORMAT_YCBCR_YUV420_3PLANE) {
			formats[num_formats++] = hvs_formats[i].drm;
		}
	}
853 854 855
	plane = &vc4_plane->base;
	ret = drm_universal_plane_init(dev, plane, 0xff,
				       &vc4_plane_funcs,
856
				       formats, num_formats,
857
				       type, NULL);
858 859 860 861 862 863 864 865 866 867

	drm_plane_helper_add(plane, &vc4_plane_helper_funcs);

	return plane;
fail:
	if (plane)
		vc4_plane_destroy(plane);

	return ERR_PTR(ret);
}