drm/rockchip: Fix YUV buffers color rendering

Currently, YUV hardware overlays are converted to RGB using
a color space conversion different than BT.601.

The result is that colors of e.g. NV12 buffers don't match
colors of YUV hardware overlays.

In order to fix this, enable YUV2YUV and set appropriate coefficients
for formats such as NV12 to be displayed correctly.

This commit was tested using modetest, gstreamer and chromeos (hardware
accelerated video playback). Before the commit, tests rendering
with NV12 format resulted in colors not displayed correctly.

Test examples (Tested on RK3399 and RK3288 boards
connected to HDMI monitor):

  $ modetest 39@32:1920x1080@NV12
  $ gst-launch-1.0 videotestrc ! video/x-raw,format=NV12 ! kmssink
Signed-off-by: Daniele Castagna <dcastagna@chromium.org>
[ezequiel: rebase on linux-next and massage commit log]
Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20190108214659.28794-1-ezequiel@collabora.com

drivers/gpu/drm/rockchip/rockchip_drm_vop.c

@@ -54,6 +54,18 @@
 		vop_reg_set(vop, &win->phy->scl->ext->name, \
 			    win->base, ~0, v, #name)
 
+#define VOP_WIN_YUV2YUV_SET(x, win_yuv2yuv, name, v) \
+	do { \
+		if (win_yuv2yuv && win_yuv2yuv->name.mask) \
+			vop_reg_set(vop, &win_yuv2yuv->name, 0, ~0, v, #name); \
+	} while (0)
+
+#define VOP_WIN_YUV2YUV_COEFFICIENT_SET(x, win_yuv2yuv, name, v) \
+	do { \
+		if (win_yuv2yuv && win_yuv2yuv->phy->name.mask) \
+			vop_reg_set(vop, &win_yuv2yuv->phy->name, win_yuv2yuv->base, ~0, v, #name); \
+	} while (0)
+
 #define VOP_INTR_SET_MASK(vop, name, mask, v) \
 		vop_reg_set(vop, &vop->data->intr->name, 0, mask, v, #name)
 
@@ -86,6 +98,18 @@
 #define to_vop(x) container_of(x, struct vop, crtc)
 #define to_vop_win(x) container_of(x, struct vop_win, base)
 
+/*
+ * The coefficients of the following matrix are all fixed points.
+ * The format is S2.10 for the 3x3 part of the matrix, and S9.12 for the offsets.
+ * They are all represented in two's complement.
+ */
+static const uint32_t bt601_yuv2rgb[] = {
+	0x4A8, 0x0,    0x662,
+	0x4A8, 0x1E6F, 0x1CBF,
+	0x4A8, 0x812,  0x0,
+	0x321168, 0x0877CF, 0x2EB127
+};
+
 enum vop_pending {
 	VOP_PENDING_FB_UNREF,
 };
@@ -93,6 +117,7 @@ enum vop_pending {
 struct vop_win {
 	struct drm_plane base;
 	const struct vop_win_data *data;
+	const struct vop_win_yuv2yuv_data *yuv2yuv_data;
 	struct vop *vop;
 };
 
@@ -714,6 +739,7 @@ static void vop_plane_atomic_update(struct drm_plane *plane,
 	struct drm_crtc *crtc = state->crtc;
 	struct vop_win *vop_win = to_vop_win(plane);
 	const struct vop_win_data *win = vop_win->data;
+	const struct vop_win_yuv2yuv_data *win_yuv2yuv = vop_win->yuv2yuv_data;
 	struct vop *vop = to_vop(state->crtc);
 	struct drm_framebuffer *fb = state->fb;
 	unsigned int actual_w, actual_h;
@@ -729,6 +755,8 @@ static void vop_plane_atomic_update(struct drm_plane *plane,
 	bool rb_swap;
 	int win_index = VOP_WIN_TO_INDEX(vop_win);
 	int format;
+	int is_yuv = fb->format->is_yuv;
+	int i;
 
 	/*
 	 * can't update plane when vop is disabled.
@@ -769,7 +797,9 @@ static void vop_plane_atomic_update(struct drm_plane *plane,
 	VOP_WIN_SET(vop, win, format, format);
 	VOP_WIN_SET(vop, win, yrgb_vir, DIV_ROUND_UP(fb->pitches[0], 4));
 	VOP_WIN_SET(vop, win, yrgb_mst, dma_addr);
-	if (fb->format->is_yuv) {
+	VOP_WIN_YUV2YUV_SET(vop, win_yuv2yuv, y2r_en, is_yuv);
+
+	if (is_yuv) {
 		int hsub = drm_format_horz_chroma_subsampling(fb->format->format);
 		int vsub = drm_format_vert_chroma_subsampling(fb->format->format);
 		int bpp = fb->format->cpp[1];
@@ -783,6 +813,13 @@ static void vop_plane_atomic_update(struct drm_plane *plane,
 		dma_addr = rk_uv_obj->dma_addr + offset + fb->offsets[1];
 		VOP_WIN_SET(vop, win, uv_vir, DIV_ROUND_UP(fb->pitches[1], 4));
 		VOP_WIN_SET(vop, win, uv_mst, dma_addr);
+
+		for (i = 0; i < NUM_YUV2YUV_COEFFICIENTS; i++) {
+			VOP_WIN_YUV2YUV_COEFFICIENT_SET(vop,
+							win_yuv2yuv,
+							y2r_coefficients[i],
+							bt601_yuv2rgb[i]);
+		}
 	}
 
 	if (win->phy->scl)
@@ -1605,6 +1642,7 @@ static void vop_win_init(struct vop *vop)
 
 		vop_win->data = win_data;
 		vop_win->vop = vop;
+		vop_win->yuv2yuv_data = &vop_data->win_yuv2yuv[i];
 	}
 }
 

drivers/gpu/drm/rockchip/rockchip_drm_vop.h

@@ -23,6 +23,8 @@
 #define VOP_MAJOR(version)		((version) >> 8)
 #define VOP_MINOR(version)		((version) & 0xff)
 
+#define NUM_YUV2YUV_COEFFICIENTS 12
+
 enum vop_data_format {
 	VOP_FMT_ARGB8888 = 0,
 	VOP_FMT_RGB888,
@@ -124,6 +126,10 @@ struct vop_scl_regs {
 	struct vop_reg scale_cbcr_y;
 };
 
+struct vop_yuv2yuv_phy {
+	struct vop_reg y2r_coefficients[NUM_YUV2YUV_COEFFICIENTS];
+};
+
 struct vop_win_phy {
 	const struct vop_scl_regs *scl;
 	const uint32_t *data_formats;
@@ -146,6 +152,12 @@ struct vop_win_phy {
 	struct vop_reg channel;
 };
 
+struct vop_win_yuv2yuv_data {
+	uint32_t base;
+	const struct vop_yuv2yuv_phy *phy;
+	struct vop_reg y2r_en;
+};
+
 struct vop_win_data {
 	uint32_t base;
 	const struct vop_win_phy *phy;
@@ -159,6 +171,7 @@ struct vop_data {
 	const struct vop_misc *misc;
 	const struct vop_modeset *modeset;
 	const struct vop_output *output;
+	const struct vop_win_yuv2yuv_data *win_yuv2yuv;
 	const struct vop_win_data *win;
 	unsigned int win_size;
 

drivers/gpu/drm/rockchip/rockchip_vop_reg.c

@@ -637,6 +637,34 @@ static const struct vop_output rk3399_output = {
 	.mipi_dual_channel_en = VOP_REG(RK3288_SYS_CTRL, 0x1, 3),
 };
 
+static const struct vop_yuv2yuv_phy rk3399_yuv2yuv_win01_data = {
+	.y2r_coefficients = {
+		VOP_REG(RK3399_WIN0_YUV2YUV_Y2R + 0, 0xffff, 0),
+		VOP_REG(RK3399_WIN0_YUV2YUV_Y2R + 0, 0xffff, 16),
+		VOP_REG(RK3399_WIN0_YUV2YUV_Y2R + 4, 0xffff, 0),
+		VOP_REG(RK3399_WIN0_YUV2YUV_Y2R + 4, 0xffff, 16),
+		VOP_REG(RK3399_WIN0_YUV2YUV_Y2R + 8, 0xffff, 0),
+		VOP_REG(RK3399_WIN0_YUV2YUV_Y2R + 8, 0xffff, 16),
+		VOP_REG(RK3399_WIN0_YUV2YUV_Y2R + 12, 0xffff, 0),
+		VOP_REG(RK3399_WIN0_YUV2YUV_Y2R + 12, 0xffff, 16),
+		VOP_REG(RK3399_WIN0_YUV2YUV_Y2R + 16, 0xffff, 0),
+		VOP_REG(RK3399_WIN0_YUV2YUV_Y2R + 20, 0xffffffff, 0),
+		VOP_REG(RK3399_WIN0_YUV2YUV_Y2R + 24, 0xffffffff, 0),
+		VOP_REG(RK3399_WIN0_YUV2YUV_Y2R + 28, 0xffffffff, 0),
+	},
+};
+
+static const struct vop_yuv2yuv_phy rk3399_yuv2yuv_win23_data = { };
+
+static const struct vop_win_yuv2yuv_data rk3399_vop_big_win_yuv2yuv_data[] = {
+	{ .base = 0x00, .phy = &rk3399_yuv2yuv_win01_data,
+	  .y2r_en = VOP_REG(RK3399_YUV2YUV_WIN, 0x1, 1) },
+	{ .base = 0x60, .phy = &rk3399_yuv2yuv_win01_data,
+	  .y2r_en = VOP_REG(RK3399_YUV2YUV_WIN, 0x1, 9) },
+	{ .base = 0xC0, .phy = &rk3399_yuv2yuv_win23_data },
+	{ .base = 0x120, .phy = &rk3399_yuv2yuv_win23_data },
+};
+
 static const struct vop_data rk3399_vop_big = {
 	.version = VOP_VERSION(3, 5),
 	.feature = VOP_FEATURE_OUTPUT_RGB10,
@@ -647,6 +675,7 @@ static const struct vop_data rk3399_vop_big = {
 	.misc = &rk3368_misc,
 	.win = rk3368_vop_win_data,
 	.win_size = ARRAY_SIZE(rk3368_vop_win_data),
+	.win_yuv2yuv = rk3399_vop_big_win_yuv2yuv_data,
 };
 
 static const struct vop_win_data rk3399_vop_lit_win_data[] = {
@@ -656,6 +685,12 @@ static const struct vop_win_data rk3399_vop_lit_win_data[] = {
 	  .type = DRM_PLANE_TYPE_CURSOR},
 };
 
+static const struct vop_win_yuv2yuv_data rk3399_vop_lit_win_yuv2yuv_data[] = {
+	{ .base = 0x00, .phy = &rk3399_yuv2yuv_win01_data,
+	  .y2r_en = VOP_REG(RK3399_YUV2YUV_WIN, 0x1, 1)},
+	{ .base = 0x60, .phy = &rk3399_yuv2yuv_win23_data },
+};
+
 static const struct vop_data rk3399_vop_lit = {
 	.version = VOP_VERSION(3, 6),
 	.intr = &rk3366_vop_intr,
@@ -665,6 +700,7 @@ static const struct vop_data rk3399_vop_lit = {
 	.misc = &rk3368_misc,
 	.win = rk3399_vop_lit_win_data,
 	.win_size = ARRAY_SIZE(rk3399_vop_lit_win_data),
+	.win_yuv2yuv = rk3399_vop_lit_win_yuv2yuv_data,
 };
 
 static const struct vop_win_data rk3228_vop_win_data[] = {