drm/vkms: Use s32 for internal color pipeline precision

Certain operations require us to preserve values below 0.0 and
above 1.0 (0x0 and 0xffff respectively in 16 bpc unorm). One
such operation is a BT709 encoding operation followed by its
decoding operation, or the reverse.

We'll use s32 values as intermediate in and outputs of our
color operations, for the operations where it matters.

For now this won't apply to LUT operations. We might want to
update those to work on s32 as well, but it's unclear how
that should work for unorm LUT definitions. We'll revisit
that once we add LUT + CTM tests.

Reviewed-by: Louis Chauvet <[e-mail]>
Signed-off-by: Alex Hung <[e-mail]>
Signed-off-by: Harry Wentland <[e-mail]>
Reviewed-by: Daniel Stone <[e-mail]>
Signed-off-by: Simon Ser <[e-mail]>
Link: https://patch.msgid.link/[e-mail]

drivers/gpu/drm/vkms/vkms_composer.c

···
 	}
 }
 
-static void apply_colorop(struct pixel_argb_u16 *pixel, struct drm_colorop *colorop)
+static void apply_colorop(struct pixel_argb_s32 *pixel, struct drm_colorop *colorop)
 {
 	struct drm_colorop_state *colorop_state = colorop->state;
 	struct drm_device *dev = colorop->dev;
···
 static void pre_blend_color_transform(const struct vkms_plane_state *plane_state,
 				      struct line_buffer *output_buffer)
 {
+	struct pixel_argb_s32 pixel;
+
 	for (size_t x = 0; x < output_buffer->n_pixels; x++) {
 		struct drm_colorop *colorop = plane_state->base.base.color_pipeline;
+
+		/*
+		 * Some operations, such as applying a BT709 encoding matrix,
+		 * followed by a decoding matrix, require that we preserve
+		 * values above 1.0 and below 0.0 until the end of the pipeline.
+		 *
+		 * Pack the 16-bit UNORM values into s32 to give us head-room to
+		 * avoid clipping until we're at the end of the pipeline. Clip
+		 * intentionally at the end of the pipeline before packing
+		 * UNORM values back into u16.
+		 */
+		pixel.a = output_buffer->pixels[x].a;
+		pixel.r = output_buffer->pixels[x].r;
+		pixel.g = output_buffer->pixels[x].g;
+		pixel.b = output_buffer->pixels[x].b;
 
 		while (colorop) {
 			struct drm_colorop_state *colorop_state;
···
 				return;
 
 			if (!colorop_state->bypass)
-				apply_colorop(&output_buffer->pixels[x], colorop);
+				apply_colorop(&pixel, colorop);
 
 			colorop = colorop->next;
 		}
+
+		/* clamp values */
+		output_buffer->pixels[x].a = clamp_val(pixel.a, 0, 0xffff);
+		output_buffer->pixels[x].r = clamp_val(pixel.r, 0, 0xffff);
+		output_buffer->pixels[x].g = clamp_val(pixel.g, 0, 0xffff);
+		output_buffer->pixels[x].b = clamp_val(pixel.b, 0, 0xffff);
 	}
 }
 

drivers/gpu/drm/vkms/vkms_drv.h

···
 	unsigned int rotation;
 };
 
+struct pixel_argb_s32 {
+	s32 a, r, g, b;
+};
+
 /**
  * struct pixel_argb_u16 - Internal representation of a pixel color.
  * @a: Alpha component value, stored in 16 bits, without padding, using