st/xa: Solid fill (tested) and composite (yet untested)

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>

10 files changed, 974 insertions, 23 deletions

diff --git a/src/gallium/state_trackers/xa/Makefile b/src/gallium/state_trackers/xa/Makefile
index 7dc47b5..92edfd0 100644
--- a/src/gallium/state_trackers/xa/Makefile
+++ b/src/gallium/state_trackers/xa/Makefile
@@ -4,7 +4,7 @@ include $(TOP)/configs/current
 ##### MACROS #####
 
 XA_MAJOR = 0
-XA_MINOR = 2
+XA_MINOR = 3
 XA_TINY = 0
 XA_CFLAGS = -g -fPIC -Wall
 
@@ -25,7 +25,8 @@ SOURCES = \
 	xa_context.c \
 	xa_renderer.c \
 	xa_tgsi.c \
-	xa_yuv.c
+	xa_yuv.c \
+	xa_composite.c
 OBJECTS = $(SOURCES:.c=.o)
 
 ##### RULES #####
diff --git a/src/gallium/state_trackers/xa/xa_composite.c b/src/gallium/state_trackers/xa/xa_composite.c
new file mode 100644
index 0000000..b307284
--- /dev/null
+++ b/src/gallium/state_trackers/xa/xa_composite.c
@@ -0,0 +1,495 @@
+/**********************************************************
+ * Copyright 2009-2011 VMware, Inc. 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, 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 AUTHORS OR COPYRIGHT HOLDERS
+ * 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.
+ *
+ *********************************************************
+ * Authors:
+ * Zack Rusin <zackr-at-vmware-dot-com>
+ * Thomas Hellstrom <thellstrom-at-vmware-dot-com>
+ */
+
+#include "xa_composite.h"
+#include "xa_context.h"
+#include "xa_priv.h"
+#include "cso_cache/cso_context.h"
+#include "util/u_sampler.h"
+#include "util/u_inlines.h"
+
+
+/*XXX also in Xrender.h but the including it here breaks compilition */
+#define XFixedToDouble(f)    (((double) (f)) / 65536.)
+
+struct xa_composite_blend {
+    enum xa_composite_op op : 8;
+
+    unsigned alpha_dst : 4;
+    unsigned alpha_src : 4;
+
+    unsigned rgb_src : 8;    /**< PIPE_BLENDFACTOR_x */
+    unsigned rgb_dst : 8;    /**< PIPE_BLENDFACTOR_x */
+};
+
+#define XA_BLEND_OP_OVER 3
+static const struct xa_composite_blend xa_blends[] = {
+    { xa_op_clear,
+      0, 0, PIPE_BLENDFACTOR_ZERO, PIPE_BLENDFACTOR_ZERO},
+    { xa_op_src,
+      0, 0, PIPE_BLENDFACTOR_ONE, PIPE_BLENDFACTOR_ZERO},
+    { xa_op_dst,
+      0, 0, PIPE_BLENDFACTOR_ZERO, PIPE_BLENDFACTOR_ONE},
+    { xa_op_over,
+      0, 1, PIPE_BLENDFACTOR_ONE, PIPE_BLENDFACTOR_INV_SRC_ALPHA},
+    { xa_op_over_reverse,
+      1, 0, PIPE_BLENDFACTOR_INV_DST_ALPHA, PIPE_BLENDFACTOR_ONE},
+    { xa_op_in,
+      1, 0, PIPE_BLENDFACTOR_DST_ALPHA, PIPE_BLENDFACTOR_ZERO},
+    { xa_op_in_reverse,
+      0, 1, PIPE_BLENDFACTOR_ZERO, PIPE_BLENDFACTOR_SRC_ALPHA},
+    { xa_op_out,
+      1, 0, PIPE_BLENDFACTOR_INV_DST_ALPHA, PIPE_BLENDFACTOR_ZERO},
+    { xa_op_out_reverse,
+      0, 1, PIPE_BLENDFACTOR_ZERO, PIPE_BLENDFACTOR_INV_SRC_ALPHA},
+    { xa_op_atop,
+      1, 1, PIPE_BLENDFACTOR_DST_ALPHA, PIPE_BLENDFACTOR_INV_SRC_ALPHA},
+    { xa_op_atop_reverse,
+      1, 1, PIPE_BLENDFACTOR_INV_DST_ALPHA, PIPE_BLENDFACTOR_SRC_ALPHA},
+    { xa_op_xor,
+      1, 1, PIPE_BLENDFACTOR_INV_DST_ALPHA, PIPE_BLENDFACTOR_INV_SRC_ALPHA},
+    { xa_op_add,
+      0, 0, PIPE_BLENDFACTOR_ONE, PIPE_BLENDFACTOR_ONE},
+};
+
+
+static boolean
+blend_for_op(struct xa_composite_blend *blend,
+	     enum xa_composite_op op,
+	     struct xa_picture *src_pic,
+	     struct xa_picture *mask_pic,
+	     struct xa_picture *dst_pic)
+{
+    const int num_blends =
+	sizeof(xa_blends)/sizeof(struct xa_composite_blend);
+    int i;
+    boolean supported = FALSE;
+
+    /*
+     * our default in case something goes wrong
+     */
+
+    *blend = xa_blends[XA_BLEND_OP_OVER];
+
+    for (i = 0; i < num_blends; ++i) {
+	if (xa_blends[i].op == op) {
+	    *blend = xa_blends[i];
+	    supported = TRUE;
+	}
+    }
+
+
+    /*
+     * If there's no dst alpha channel, adjust the blend op so that we'll treat
+     * it as always 1.
+     */
+
+    if (dst_pic &&
+	xa_format_a(xa_surface_format(dst_pic->srf)) == 0 &&
+	blend->alpha_dst) {
+	if (blend->rgb_src == PIPE_BLENDFACTOR_DST_ALPHA)
+	    blend->rgb_src = PIPE_BLENDFACTOR_ONE;
+	else if (blend->rgb_src == PIPE_BLENDFACTOR_INV_DST_ALPHA)
+	    blend->rgb_src = PIPE_BLENDFACTOR_ZERO;
+    }
+
+    /*
+     * If the source alpha is being used, then we should only be in a case where
+     * the source blend factor is 0, and the source blend value is the mask
+     * channels multiplied by the source picture's alpha.
+     */
+    if (mask_pic && mask_pic->component_alpha &&
+	xa_format_rgb(xa_surface_format(mask_pic->srf)) &&
+	blend->alpha_src) {
+	if (blend->rgb_dst == PIPE_BLENDFACTOR_SRC_ALPHA) {
+	    blend->rgb_dst = PIPE_BLENDFACTOR_SRC_COLOR;
+	} else if (blend->rgb_dst == PIPE_BLENDFACTOR_INV_SRC_ALPHA) {
+	    blend->rgb_dst = PIPE_BLENDFACTOR_INV_SRC_COLOR;
+	}
+    }
+
+    return supported;
+}
+
+
+static INLINE int
+xa_repeat_to_gallium(int mode)
+{
+    switch(mode) {
+    case xa_wrap_clamp_to_border:
+	return PIPE_TEX_WRAP_CLAMP_TO_BORDER;
+    case xa_wrap_repeat:
+	return PIPE_TEX_WRAP_REPEAT;
+    case xa_wrap_mirror_repeat:
+	return PIPE_TEX_WRAP_MIRROR_REPEAT;
+    case xa_wrap_clamp_to_edge:
+	return PIPE_TEX_WRAP_CLAMP_TO_EDGE;
+    default:
+	break;
+    }
+    return PIPE_TEX_WRAP_REPEAT;
+}
+
+static INLINE boolean
+xa_filter_to_gallium(int xrender_filter, int *out_filter)
+{
+
+    switch (xrender_filter) {
+    case xa_filter_nearest:
+	*out_filter = PIPE_TEX_FILTER_NEAREST;
+	break;
+    case xa_filter_linear:
+	*out_filter = PIPE_TEX_FILTER_LINEAR;
+	break;
+    default:
+	*out_filter = PIPE_TEX_FILTER_NEAREST;
+	return FALSE;
+    }
+    return TRUE;
+}
+
+static int
+xa_is_filter_accelerated(struct xa_picture *pic)
+{
+    int filter;
+    if (pic && !xa_filter_to_gallium(pic->filter, &filter))
+	return 0;
+    return 1;
+}
+
+int
+xa_composite_is_accelerated(const struct xa_composite *comp)
+{
+    struct xa_composite_blend blend;
+    struct xa_picture *src_pic = comp->src;
+
+    if (!xa_is_filter_accelerated(src_pic) ||
+	!xa_is_filter_accelerated(comp->mask)) {
+	return 0;
+    }
+
+
+    if (src_pic->src_pict) {
+	if (src_pic->src_pict->type != xa_src_pict_solid_fill)
+	    return 0;
+    }
+
+    if (blend_for_op(&blend, comp->op, comp->src, comp->mask, comp->dst)) {
+	struct xa_picture *mask = comp->mask;
+	if (mask && mask->component_alpha &&
+	    xa_format_rgb(xa_surface_format(mask->srf))) {
+	    if (blend.alpha_src && blend.rgb_src != PIPE_BLENDFACTOR_ZERO) {
+		return 0;
+	    }
+	}
+
+	return 1;
+    }
+    return 0;
+}
+
+static void
+bind_composite_blend_state(struct xa_context *ctx,
+			   const struct xa_composite *comp)
+{
+    struct xa_composite_blend blend_opt;
+    struct pipe_blend_state blend;
+
+    blend_for_op(&blend_opt, comp->op, comp->src, comp->mask, comp->dst);
+
+    memset(&blend, 0, sizeof(struct pipe_blend_state));
+    blend.rt[0].blend_enable = 1;
+    blend.rt[0].colormask = PIPE_MASK_RGBA;
+
+    blend.rt[0].rgb_src_factor   = blend_opt.rgb_src;
+    blend.rt[0].alpha_src_factor = blend_opt.rgb_src;
+    blend.rt[0].rgb_dst_factor   = blend_opt.rgb_dst;
+    blend.rt[0].alpha_dst_factor = blend_opt.rgb_dst;
+
+    cso_set_blend(ctx->cso, &blend);
+}
+
+static unsigned int
+picture_format_fixups(struct xa_picture *src_pic,
+		      struct xa_picture *dst_pic,
+		      int mask)
+{
+    boolean set_alpha = FALSE;
+    boolean swizzle = FALSE;
+    unsigned ret = 0;
+    struct xa_surface *src = src_pic->srf;
+    enum xa_formats src_hw_format, src_pic_format;
+    enum xa_surface_type src_hw_type, src_pic_type;
+
+    src_hw_format = xa_surface_format(src);
+    src_pic_format = src_pic->pict_format;
+
+    if (src_hw_format == src_pic_format) {
+	if (src_pic->pict_format == xa_format_a8) {
+	    if (mask)
+		return FS_MASK_LUMINANCE;
+	    else if (dst_pic->pict_format != xa_format_a8) {
+
+		/*
+		 * if both dst and src are luminance then
+		 * we don't want to swizzle the alpha (X) of the
+		 * source into W component of the dst because
+		 * it will break our destination
+		 */
+		return FS_SRC_LUMINANCE;
+	    }
+	}
+	return 0;
+    }
+
+    src_hw_type = xa_format_type(src_hw_format);
+    src_pic_type = xa_format_type(src_pic_format);
+
+    swizzle = ((src_hw_type == xa_type_argb &&
+		src_pic_type == xa_type_abgr) ||
+	       ((src_hw_type == xa_type_abgr &&
+		 src_pic_type == xa_type_argb)));
+
+    if (!swizzle && (src_hw_type != src_pic_type))
+	return 0;
+
+    set_alpha = (xa_format_type_is_color(src_hw_type) &&
+		 xa_format_a(src_hw_type) == 0);
+
+    if (set_alpha)
+	ret |= mask ? FS_MASK_SET_ALPHA : FS_SRC_SET_ALPHA;
+    if (swizzle)
+	ret |= mask ? FS_MASK_SWIZZLE_RGB : FS_SRC_SWIZZLE_RGB;
+
+    return ret;
+}
+
+static void
+bind_shaders(struct xa_context *ctx, const struct xa_composite *comp)
+{
+    unsigned vs_traits = 0, fs_traits = 0;
+    struct xa_shader shader;
+    struct xa_picture *src_pic = comp->src;
+    struct xa_picture *mask_pic = comp->mask;
+    struct xa_picture *dst_pic = comp->dst;
+
+    ctx->has_solid_color = FALSE;
+
+    if (src_pic) {
+	if (src_pic->wrap == xa_wrap_clamp_to_border && src_pic->has_transform)
+	    fs_traits |= FS_SRC_REPEAT_NONE;
+
+	if (src_pic->src_pict) {
+	    if (src_pic->src_pict->type == xa_src_pict_solid_fill) {
+		fs_traits |= FS_SOLID_FILL;
+		vs_traits |= VS_SOLID_FILL;
+		xa_pixel_to_float4(src_pic->src_pict->solid_fill.color,
+				   ctx->solid_color);
+		ctx->has_solid_color = TRUE;
+	    }
+	} else {
+	    fs_traits |= FS_COMPOSITE;
+	    vs_traits |= VS_COMPOSITE;
+	}
+
+	fs_traits |= picture_format_fixups(src_pic, dst_pic, 0);
+    }
+
+    if (mask_pic) {
+	vs_traits |= VS_MASK;
+	fs_traits |= FS_MASK;
+	if (mask_pic->wrap == xa_wrap_clamp_to_border &&
+	    mask_pic->has_transform)
+	    fs_traits |= FS_MASK_REPEAT_NONE;
+
+	if (mask_pic->component_alpha) {
+	    struct xa_composite_blend blend;
+	    blend_for_op(&blend, comp->op, src_pic, mask_pic, NULL);
+	    if (blend.alpha_src) {
+		fs_traits |= FS_CA_SRCALPHA;
+	    } else
+		fs_traits |= FS_CA_FULL;
+	}
+
+	fs_traits |= picture_format_fixups(mask_pic, dst_pic, 1);
+    }
+
+    shader = xa_shaders_get(ctx->shaders, vs_traits, fs_traits);
+    cso_set_vertex_shader_handle(ctx->cso, shader.vs);
+    cso_set_fragment_shader_handle(ctx->cso, shader.fs);
+}
+
+static void
+bind_samplers(struct xa_context *ctx,
+	      const struct xa_composite *comp)
+{
+    struct pipe_sampler_state *samplers[PIPE_MAX_SAMPLERS];
+    struct pipe_sampler_state src_sampler, mask_sampler;
+    struct pipe_sampler_view view_templ;
+    struct pipe_sampler_view *src_view;
+    struct pipe_context *pipe = ctx->pipe;
+    struct xa_picture *src_pic = comp->src;
+    struct xa_picture *mask_pic = comp->mask;
+
+    ctx->num_bound_samplers = 0;
+
+    memset(&src_sampler, 0, sizeof(struct pipe_sampler_state));
+    memset(&mask_sampler, 0, sizeof(struct pipe_sampler_state));
+
+    if (src_pic) {
+	if (ctx->has_solid_color) {
+	    samplers[0] = NULL;
+	    pipe_sampler_view_reference(&ctx->bound_sampler_views[0], NULL);
+	} else {
+	    unsigned src_wrap = xa_repeat_to_gallium(src_pic->wrap);
+	    int filter;
+
+	    (void) xa_filter_to_gallium(src_pic->filter, &filter);
+
+	    src_sampler.wrap_s = src_wrap;
+	    src_sampler.wrap_t = src_wrap;
+	    src_sampler.min_img_filter = filter;
+	    src_sampler.mag_img_filter = filter;
+	    src_sampler.min_mip_filter = PIPE_TEX_MIPFILTER_NEAREST;
+	    src_sampler.normalized_coords = 1;
+	    samplers[0] = &src_sampler;
+	    ctx->num_bound_samplers = 1;
+	    u_sampler_view_default_template(&view_templ,
+					    src_pic->srf->tex,
+					    src_pic->srf->tex->format);
+	    src_view = pipe->create_sampler_view(pipe, src_pic->srf->tex,
+						 &view_templ);
+	    pipe_sampler_view_reference(&ctx->bound_sampler_views[0], NULL);
+	    ctx->bound_sampler_views[0] = src_view;
+	}
+    }
+
+    if (mask_pic) {
+	unsigned mask_wrap = xa_repeat_to_gallium(mask_pic->wrap);
+	int filter;
+
+	(void) xa_filter_to_gallium(mask_pic->filter, &filter);
+
+	mask_sampler.wrap_s = mask_wrap;
+	mask_sampler.wrap_t = mask_wrap;
+	mask_sampler.min_img_filter = filter;
+	mask_sampler.mag_img_filter = filter;
+	src_sampler.min_mip_filter = PIPE_TEX_MIPFILTER_NEAREST;
+	mask_sampler.normalized_coords = 1;
+	samplers[1] = &mask_sampler;
+	ctx->num_bound_samplers = 2;
+	u_sampler_view_default_template(&view_templ,
+					mask_pic->srf->tex,
+					mask_pic->srf->tex->format);
+	src_view = pipe->create_sampler_view(pipe, mask_pic->srf->tex,
+					     &view_templ);
+	pipe_sampler_view_reference(&ctx->bound_sampler_views[1], NULL);
+	ctx->bound_sampler_views[1] = src_view;
+    }
+
+    cso_set_samplers(ctx->cso, ctx->num_bound_samplers,
+		     (const struct pipe_sampler_state **)samplers);
+    cso_set_fragment_sampler_views(ctx->cso, ctx->num_bound_samplers,
+				   ctx->bound_sampler_views);
+}
+
+int
+xa_composite_prepare(struct xa_context *ctx,
+		     const struct xa_composite *comp)
+{
+    struct xa_surface *dst_srf = comp->dst->srf;
+    int ret;
+
+    ret = xa_surface_psurf_create(ctx, dst_srf);
+    if (ret != XA_ERR_NONE)
+	return ret;
+
+    renderer_bind_destination(ctx, dst_srf->srf,
+			      dst_srf->srf->width,
+			      dst_srf->srf->height);
+
+    bind_composite_blend_state(ctx, comp);
+    bind_shaders(ctx, comp);
+    bind_samplers(ctx, comp);
+
+    if (ctx->num_bound_samplers == 0 ) { /* solid fill */
+	renderer_begin_solid(ctx);
+    } else {
+	renderer_begin_textures(ctx);
+    }
+
+    xa_surface_psurf_destroy(dst_srf);
+    return XA_ERR_NONE;
+}
+
+void xa_composite_rect(struct xa_context *ctx,
+		       int srcX, int srcY, int maskX, int maskY,
+		       int dstX, int dstY, int width, int height)
+{
+    if (ctx->num_bound_samplers == 0 ) { /* solid fill */
+	renderer_solid(ctx, dstX, dstY, dstX + width, dstY + height,
+		       ctx->solid_color);
+    } else {
+	const struct xa_composite *comp = ctx->comp;
+	int pos[6] = {srcX, srcY, maskX, maskY, dstX, dstY};
+	const float *src_matrix = NULL;
+	const float *mask_matrix = NULL;
+
+	if (comp->src->has_transform)
+	    src_matrix = comp->src->transform;
+	if (comp->mask->has_transform)
+	    mask_matrix = comp->mask->transform;
+
+	renderer_texture(ctx, pos, width, height,
+			 src_matrix, mask_matrix);
+    }
+}
+
+void
+xa_composite_done(struct xa_context *ctx)
+{
+    renderer_draw_flush(ctx);
+    ctx->pipe->flush(ctx->pipe, &ctx->last_fence);
+
+    ctx->comp = NULL;
+    ctx->has_solid_color = FALSE;
+    ctx->num_bound_samplers = 0;
+}
+
+static const struct xa_composite_allocation a = {
+    .xa_composite_size = sizeof(struct xa_composite),
+    .xa_picture_size = sizeof(struct xa_picture),
+    .xa_source_pict_size = sizeof(union xa_source_pict),
+};
+
+const struct xa_composite_allocation *
+xa_composite_allocation(void)
+{
+    return &a;
+}
diff --git a/src/gallium/state_trackers/xa/xa_composite.h b/src/gallium/state_trackers/xa/xa_composite.h
new file mode 100644
index 0000000..63174a4
--- /dev/null
+++ b/src/gallium/state_trackers/xa/xa_composite.h
@@ -0,0 +1,131 @@
+/**********************************************************
+ * Copyright 2009-2011 VMware, Inc. 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, 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 AUTHORS OR COPYRIGHT HOLDERS
+ * 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.
+ *
+ *********************************************************
+ * Authors:
+ * Zack Rusin <zackr-at-vmware-dot-com>
+ * Thomas Hellstrom <thellstrom-at-vmware-dot-com>
+ */
+
+#ifndef _XA_COMPOSITE_H_
+#define _XA_COMPOSITE_H_
+
+#include "xa_tracker.h"
+#include "xa_context.h"
+
+/*
+ * Supported composite ops.
+ */
+enum xa_composite_op {
+    xa_op_clear,
+    xa_op_src,
+    xa_op_dst,
+    xa_op_over,
+    xa_op_over_reverse,
+    xa_op_in,
+    xa_op_in_reverse,
+    xa_op_out,
+    xa_op_out_reverse,
+    xa_op_atop,
+    xa_op_atop_reverse,
+    xa_op_xor,
+    xa_op_add
+};
+
+/*
+ * Supported filters.
+ */
+enum xa_composite_filter {
+    xa_filter_nearest,
+    xa_filter_linear
+};
+
+/*
+ * Supported clamp methods.
+ */
+enum xa_composite_wrap {
+    xa_wrap_clamp_to_border,
+    xa_wrap_repeat,
+    xa_wrap_mirror_repeat,
+    xa_wrap_clamp_to_edge
+};
+
+/*
+ * Src picture types.
+ */
+enum xa_composite_src_pict_type {
+    xa_src_pict_solid_fill
+};
+
+struct xa_pict_solid_fill {
+    enum xa_composite_src_pict_type type;
+    unsigned int class;
+    uint32_t color;
+};
+
+union xa_source_pict {
+    unsigned int type;
+    struct xa_pict_solid_fill solid_fill;
+};
+
+struct xa_picture {
+    enum xa_formats pict_format;
+    struct xa_surface *srf;
+    struct xa_surface *alpha_map;
+    float transform[9];
+    int has_transform;
+    int component_alpha;
+    enum xa_composite_wrap wrap;
+    enum xa_composite_filter filter;
+    union xa_source_pict *src_pict;
+};
+
+struct xa_composite {
+    struct xa_picture *src, *mask, *dst;
+    int op;
+};
+
+struct xa_composite_allocation {
+    unsigned int xa_composite_size;
+    unsigned int xa_picture_size;
+    unsigned int xa_source_pict_size;
+};
+
+/*
+ * Get allocation sizes for minor bump compatibility.
+ */
+
+extern const struct xa_composite_allocation *
+xa_composite_allocation(void);
+extern int
+xa_composite_is_accelerated(const struct xa_composite *comp);
+extern int
+xa_composite_prepare(struct xa_context *ctx, const struct xa_composite *comp);
+extern void
+xa_composite_rect(struct xa_context *ctx,
+		  int srcX, int srcY, int maskX, int maskY,
+		  int dstX, int dstY, int width, int height);
+extern void
+xa_composite_done(struct xa_context *ctx);
+
+#endif
diff --git a/src/gallium/state_trackers/xa/xa_context.c b/src/gallium/state_trackers/xa/xa_context.c
index a6abd8c..3cc25ed 100644
--- a/src/gallium/state_trackers/xa/xa_context.c
+++ b/src/gallium/state_trackers/xa/xa_context.c
@@ -174,6 +174,35 @@ xa_surface_unmap(struct xa_surface *srf)
 }
 
 int
+xa_surface_psurf_create(struct xa_context *ctx, struct xa_surface *dst)
+{
+    struct pipe_screen *screen = ctx->pipe->screen;
+    struct pipe_surface srf_templ;
+
+    if (dst->srf)
+	return -XA_ERR_INVAL;
+
+    if (!screen->is_format_supported(screen,  dst->tex->format,
+				     PIPE_TEXTURE_2D, 0,
+				     PIPE_BIND_RENDER_TARGET))
+	return -XA_ERR_INVAL;
+
+    u_surface_default_template(&srf_templ, dst->tex,
+			       PIPE_BIND_RENDER_TARGET);
+    dst->srf = ctx->pipe->create_surface(ctx->pipe, dst->tex, &srf_templ);
+    if (!dst->srf)
+	return -XA_ERR_NORES;
+
+    return XA_ERR_NONE;
+}
+
+void
+xa_surface_psurf_destroy(struct xa_surface *dst)
+{
+    pipe_surface_reference(&dst->srf, NULL);
+}
+
+int
 xa_copy_prepare(struct xa_context *ctx,
 		struct xa_surface *dst, struct xa_surface *src)
 {
@@ -181,19 +210,9 @@ xa_copy_prepare(struct xa_context *ctx,
 	return -XA_ERR_INVAL;
 
     if (src->tex->format != dst->tex->format) {
-	struct pipe_screen *screen = ctx->pipe->screen;
-	struct pipe_surface srf_templ;
-
-	if (!screen->is_format_supported(screen,  dst->tex->format,
-					 PIPE_TEXTURE_2D, 0,
-					 PIPE_BIND_RENDER_TARGET))
-	    return -XA_ERR_INVAL;
-	u_surface_default_template(&srf_templ, dst->tex,
-				   PIPE_BIND_RENDER_TARGET);
-	dst->srf = ctx->pipe->create_surface(ctx->pipe, dst->tex, &srf_templ);
-	if (!dst->srf)
-	    return -XA_ERR_NORES;
-
+	int ret = xa_surface_psurf_create(ctx, dst);
+	if (ret != XA_ERR_NONE)
+	    return ret;
 	renderer_copy_prepare(ctx, dst->srf, src->tex);
 	ctx->simple_copy = 0;
     } else
@@ -228,11 +247,91 @@ xa_copy_done(struct xa_context *ctx)
     if (!ctx->simple_copy) {
 	   renderer_draw_flush(ctx);
 	   ctx->pipe->flush(ctx->pipe, &ctx->last_fence);
-	   pipe_surface_reference(&ctx->dst->srf, NULL);
+	   xa_surface_psurf_destroy(ctx->dst);
     } else
 	ctx->pipe->flush(ctx->pipe, &ctx->last_fence);
 }
 
+static void
+bind_solid_blend_state(struct xa_context *ctx)
+{
+    struct pipe_blend_state blend;
+
+    memset(&blend, 0, sizeof(struct pipe_blend_state));
+    blend.rt[0].blend_enable = 0;
+    blend.rt[0].colormask = PIPE_MASK_RGBA;
+
+    blend.rt[0].rgb_src_factor   = PIPE_BLENDFACTOR_ONE;
+    blend.rt[0].alpha_src_factor = PIPE_BLENDFACTOR_ONE;
+    blend.rt[0].rgb_dst_factor   = PIPE_BLENDFACTOR_ZERO;
+    blend.rt[0].alpha_dst_factor = PIPE_BLENDFACTOR_ZERO;
+
+    cso_set_blend(ctx->cso, &blend);
+}
+
+int
+xa_solid_prepare(struct xa_context *ctx, struct xa_surface *dst,
+		 uint32_t fg)
+{
+    unsigned vs_traits, fs_traits;
+    struct xa_shader shader;
+    int width, height;
+    int ret;
+
+    xa_pixel_to_float4(fg, ctx->solid_color);
+    ctx->has_solid_color = 1;
+
+    ret = xa_surface_psurf_create(ctx, dst);
+    if (ret != XA_ERR_NONE)
+	return ret;
+
+    ctx->dst = dst;
+    width = dst->srf->width;
+    height = dst->srf->height;
+
+#if 0
+    debug_printf("Color Pixel=(%d, %d, %d, %d), RGBA=(%f, %f, %f, %f)\n",
+		 (fg >> 24) & 0xff, (fg >> 16) & 0xff,
+		 (fg >> 8) & 0xff,  (fg >> 0) & 0xff,
+		 exa->solid_color[0], exa->solid_color[1],
+		 exa->solid_color[2], exa->solid_color[3]);
+#endif
+
+    vs_traits = VS_SOLID_FILL;
+    fs_traits = FS_SOLID_FILL;
+
+    renderer_bind_destination(ctx, dst->srf, width, height);
+    bind_solid_blend_state(ctx);
+    cso_set_samplers(ctx->cso, 0, NULL);
+    cso_set_fragment_sampler_views(ctx->cso, 0, NULL);
+
+    shader = xa_shaders_get(ctx->shaders, vs_traits, fs_traits);
+    cso_set_vertex_shader_handle(ctx->cso, shader.vs);
+    cso_set_fragment_shader_handle(ctx->cso, shader.fs);
+
+    renderer_begin_solid(ctx);
+
+    xa_surface_psurf_destroy(dst);
+    return XA_ERR_NONE;
+}
+
+void
+xa_solid(struct xa_context *ctx, int x, int y, int width, int height)
+{
+    renderer_solid(ctx, x, y, x + width, y + height, ctx->solid_color);
+}
+
+void
+xa_solid_done(struct xa_context *ctx)
+{
+    renderer_draw_flush(ctx);
+    ctx->pipe->flush(ctx->pipe, &ctx->last_fence);
+
+    ctx->comp = NULL;
+    ctx->has_solid_color = FALSE;
+    ctx->num_bound_samplers = 0;
+}
+
 struct xa_fence *
 xa_fence_get(struct xa_context *ctx)
 {
diff --git a/src/gallium/state_trackers/xa/xa_context.h b/src/gallium/state_trackers/xa/xa_context.h
index 753ba3d..ea2b923 100644
--- a/src/gallium/state_trackers/xa/xa_context.h
+++ b/src/gallium/state_trackers/xa/xa_context.h
@@ -69,6 +69,15 @@ extern int xa_surface_dma(struct xa_context *ctx,
 			  int to_surface, struct xa_box *boxes,
 			  unsigned int num_boxes);
 
+extern int
+xa_solid_prepare(struct xa_context *ctx, struct xa_surface *dst,
+		 uint32_t fg);
+extern void
+xa_solid(struct xa_context *ctx, int x, int y, int width, int height);
+
+extern void
+xa_solid_done(struct xa_context *ctx);
+
 extern struct xa_fence *xa_fence_get(struct xa_context *ctx);
 
 extern int xa_fence_wait(struct xa_fence *fence, uint64_t timeout);
diff --git a/src/gallium/state_trackers/xa/xa_priv.h b/src/gallium/state_trackers/xa/xa_priv.h
index 22a49c1..94627e1 100644
--- a/src/gallium/state_trackers/xa/xa_priv.h
+++ b/src/gallium/state_trackers/xa/xa_priv.h
@@ -32,6 +32,7 @@
 
 #include "xa_tracker.h"
 #include "xa_context.h"
+#include "xa_composite.h"
 
 #include "pipe/p_screen.h"
 #include "pipe/p_context.h"
@@ -39,6 +40,7 @@
 
 #define XA_VB_SIZE (100 * 4 * 3 * 4)
 #define XA_LAST_SURFACE_TYPE (xa_type_yuv_component + 1)
+#define XA_MAX_SAMPLERS 3
 
 struct xa_fence {
     struct pipe_fence_handle *pipe_fence;
@@ -96,6 +98,13 @@ struct xa_context {
     struct xa_surface *src;
     struct xa_surface *dst;
     int simple_copy;
+
+    int has_solid_color;
+    float solid_color[4];
+
+    unsigned int num_bound_samplers;
+    struct pipe_sampler_view *bound_sampler_views[XA_MAX_SAMPLERS];
+    const struct xa_composite *comp;
 };
 
 enum xa_vs_traits {
@@ -136,13 +145,33 @@ struct xa_shader {
     void *vs;
 };
 
-static inline int
+struct xa_shaders;
+
+/*
+ * Inline utilities
+ */
+
+static INLINE int
 xa_min(int a, int b)
 {
     return ((a <= b) ? a : b);
 }
 
-struct xa_shaders;
+static INLINE void
+xa_pixel_to_float4(uint32_t pixel, float *color)
+{
+    uint32_t	    r, g, b, a;
+
+    a = (pixel >> 24) & 0xff;
+    r = (pixel >> 16) & 0xff;
+    g = (pixel >>  8) & 0xff;
+    b = (pixel >>  0) & 0xff;
+    color[0] = ((float)r) / 255.;
+    color[1] = ((float)g) / 255.;
+    color[2] = ((float)b) / 255.;
+    color[3] = ((float)a) / 255.;
+}
+
 
 /*
  * xa_tgsi.c
@@ -156,6 +185,15 @@ struct xa_shader xa_shaders_get(struct xa_shaders *shaders,
 				unsigned vs_traits, unsigned fs_traits);
 
 /*
+ * xa_context.c
+ */
+extern int
+xa_surface_psurf_create(struct xa_context *ctx, struct xa_surface *dst);
+
+extern void
+xa_surface_psurf_destroy(struct xa_surface *dst);
+
+/*
  * xa_renderer.c
  */
 void renderer_set_constants(struct xa_context *r,
@@ -187,4 +225,17 @@ void renderer_copy(struct xa_context *r, int dx,
 
 void renderer_draw_flush(struct xa_context *r);
 
+void renderer_begin_solid(struct xa_context *r);
+void renderer_solid(struct xa_context *r,
+		    int x0, int y0, int x1, int y1, float *color);
+void
+renderer_begin_textures(struct xa_context *r);
+
+void
+renderer_texture(struct xa_context *r,
+		 int *pos,
+		 int width, int height,
+		 const float *src_matrix,
+		 const float *mask_matrix);
+
 #endif
diff --git a/src/gallium/state_trackers/xa/xa_renderer.c b/src/gallium/state_trackers/xa/xa_renderer.c
index b92df0d..559b269 100644
--- a/src/gallium/state_trackers/xa/xa_renderer.c
+++ b/src/gallium/state_trackers/xa/xa_renderer.c
@@ -202,6 +202,107 @@ add_vertex_2tex(struct xa_context *r,
     r->buffer_size += 12;
 }
 
+static void
+add_vertex_data1(struct xa_context *r,
+                 float srcX, float srcY,  float dstX, float dstY,
+                 float width, float height,
+                 struct pipe_resource *src, const float *src_matrix)
+{
+    float s0, t0, s1, t1, s2, t2, s3, t3;
+    float pt0[2], pt1[2], pt2[2], pt3[2];
+
+    pt0[0] = srcX;
+    pt0[1] = srcY;
+    pt1[0] = (srcX + width);
+    pt1[1] = srcY;
+    pt2[0] = (srcX + width);
+    pt2[1] = (srcY + height);
+    pt3[0] = srcX;
+    pt3[1] = (srcY + height);
+
+    if (src_matrix) {
+	map_point((float *)src_matrix, pt0[0], pt0[1], &pt0[0], &pt0[1]);
+	map_point((float *)src_matrix, pt1[0], pt1[1], &pt1[0], &pt1[1]);
+	map_point((float *)src_matrix, pt2[0], pt2[1], &pt2[0], &pt2[1]);
+	map_point((float *)src_matrix, pt3[0], pt3[1], &pt3[0], &pt3[1]);
+    }
+
+    s0 =  pt0[0] / src->width0;
+    s1 =  pt1[0] / src->width0;
+    s2 =  pt2[0] / src->width0;
+    s3 =  pt3[0] / src->width0;
+    t0 =  pt0[1] / src->height0;
+    t1 =  pt1[1] / src->height0;
+    t2 =  pt2[1] / src->height0;
+    t3 =  pt3[1] / src->height0;
+
+    /* 1st vertex */
+    add_vertex_1tex(r, dstX, dstY, s0, t0);
+    /* 2nd vertex */
+    add_vertex_1tex(r, dstX + width, dstY, s1, t1);
+    /* 3rd vertex */
+    add_vertex_1tex(r, dstX + width, dstY + height, s2, t2);
+    /* 4th vertex */
+    add_vertex_1tex(r, dstX, dstY + height, s3, t3);
+}
+
+static void
+add_vertex_data2(struct xa_context *r,
+                 float srcX, float srcY, float maskX, float maskY,
+                 float dstX, float dstY, float width, float height,
+                 struct pipe_resource *src,
+                 struct pipe_resource *mask,
+                 const float *src_matrix, const float *mask_matrix)
+{
+    float src_s0, src_t0, src_s1, src_t1;
+    float mask_s0, mask_t0, mask_s1, mask_t1;
+    float spt0[2], spt1[2];
+    float mpt0[2], mpt1[2];
+
+    spt0[0] = srcX;
+    spt0[1] = srcY;
+    spt1[0] = srcX + width;
+    spt1[1] = srcY + height;
+
+    mpt0[0] = maskX;
+    mpt0[1] = maskY;
+    mpt1[0] = maskX + width;
+    mpt1[1] = maskY + height;
+
+    if (src_matrix) {
+	map_point((float *)src_matrix, spt0[0], spt0[1], &spt0[0], &spt0[1]);
+	map_point((float *)src_matrix, spt1[0], spt1[1], &spt1[0], &spt1[1]);
+    }
+
+    if (mask_matrix) {
+	map_point((float *)mask_matrix, mpt0[0], mpt0[1], &mpt0[0], &mpt0[1]);
+	map_point((float *)mask_matrix, mpt1[0], mpt1[1], &mpt1[0], &mpt1[1]);
+    }
+
+    src_s0 = spt0[0] / src->width0;
+    src_t0 = spt0[1] / src->height0;
+    src_s1 = spt1[0] / src->width0;
+    src_t1 = spt1[1] / src->height0;
+
+    mask_s0 = mpt0[0] / mask->width0;
+    mask_t0 = mpt0[1] / mask->height0;
+    mask_s1 = mpt1[0] / mask->width0;
+    mask_t1 = mpt1[1] / mask->height0;
+
+    /* 1st vertex */
+    add_vertex_2tex(r, dstX, dstY,
+		    src_s0, src_t0, mask_s0, mask_t0);
+    /* 2nd vertex */
+    add_vertex_2tex(r, dstX + width, dstY,
+		    src_s1, src_t0, mask_s1, mask_t0);
+    /* 3rd vertex */
+    add_vertex_2tex(r, dstX + width, dstY + height,
+		    src_s1, src_t1, mask_s1, mask_t1);
+    /* 4th vertex */
+    add_vertex_2tex(r, dstX, dstY + height,
+		    src_s0, src_t1, mask_s0, mask_t1);
+}
+
 static struct pipe_resource *
 setup_vertex_data_yuv(struct xa_context *r,
 		      float srcX,
@@ -468,3 +569,58 @@ renderer_draw_flush(struct xa_context *r)
 {
     renderer_draw_conditional(r, 0);
 }
+
+void
+renderer_begin_textures(struct xa_context *r)
+{
+    r->attrs_per_vertex = 1 + r->num_bound_samplers;
+    r->buffer_size = 0;
+}
+
+void
+renderer_texture(struct xa_context *r,
+		 int *pos,
+		 int width, int height,
+		 const float *src_matrix,
+		 const float *mask_matrix)
+{
+    struct pipe_sampler_view **sampler_view = r->bound_sampler_views;
+
+#if 0
+    if (src_matrix) {
+	debug_printf("src_matrix = \n");
+	debug_printf("%f, %f, %f\n", src_matrix[0], src_matrix[1], src_matrix[2]);
+	debug_printf("%f, %f, %f\n", src_matrix[3], src_matrix[4], src_matrix[5]);
+	debug_printf("%f, %f, %f\n", src_matrix[6], src_matrix[7], src_matrix[8]);
+    }
+    if (mask_matrix) {
+	debug_printf("mask_matrix = \n");
+	debug_printf("%f, %f, %f\n", mask_matrix[0], mask_matrix[1], mask_matrix[2]);
+	debug_printf("%f, %f, %f\n", mask_matrix[3], mask_matrix[4], mask_matrix[5]);
+	debug_printf("%f, %f, %f\n", mask_matrix[6], mask_matrix[7], mask_matrix[8]);
+    }
+#endif
+
+    switch(r->attrs_per_vertex) {
+    case 2:
+	renderer_draw_conditional(r, 4 * 8);
+	add_vertex_data1(r,
+			 pos[0], pos[1], /* src */
+			 pos[4], pos[5], /* dst */
+			 width, height,
+			 sampler_view[0]->texture, src_matrix);
+	break;
+    case 3:
+	renderer_draw_conditional(r, 4 * 12);
+	add_vertex_data2(r,
+			 pos[0], pos[1], /* src */
+			 pos[2], pos[3], /* mask */
+			 pos[4], pos[5], /* dst */
+			 width, height,
+			 sampler_view[0]->texture, sampler_view[1]->texture,
+			 src_matrix, mask_matrix);
+	break;
+    default:
+	break;
+    }
+}
diff --git a/src/gallium/state_trackers/xa/xa_symbols b/src/gallium/state_trackers/xa/xa_symbols
index 62ae333..7fa6c94 100644
--- a/src/gallium/state_trackers/xa/xa_symbols
+++ b/src/gallium/state_trackers/xa/xa_symbols
@@ -8,9 +8,6 @@ xa_surface_dma
 xa_surface_map
 xa_surface_unmap
 xa_surface_format
-xa_copy_prepare
-xa_copy
-xa_copy_done
 xa_surface_handle
 xa_context_default
 xa_context_create
@@ -18,4 +15,15 @@ xa_context_destroy
 xa_fence_get
 xa_fence_wait
 xa_fence_destroy
+xa_copy_prepare
+xa_copy
+xa_copy_done
+xa_solid_prepare
+xa_solid
+xa_solid_done
+xa_composite_allocation
+xa_composite_is_accelerated
+xa_composite_prepare
+xa_composite_rect
+xa_composite_done
 xa_yuv_planar_blit
diff --git a/src/gallium/state_trackers/xa/xa_tracker.h b/src/gallium/state_trackers/xa/xa_tracker.h
index ed81070..8c9dd5e 100644
--- a/src/gallium/state_trackers/xa/xa_tracker.h
+++ b/src/gallium/state_trackers/xa/xa_tracker.h
@@ -37,7 +37,7 @@
 #include <stdint.h>
 
 #define XA_TRACKER_VERSION_MAJOR 0
-#define XA_TRACKER_VERSION_MINOR 2
+#define XA_TRACKER_VERSION_MINOR 3
 #define XA_TRACKER_VERSION_PATCH 0
 
 #define XA_FLAG_SHARED         (1 << 0)
diff --git a/src/gallium/targets/xa-vmwgfx/Makefile b/src/gallium/targets/xa-vmwgfx/Makefile
index 7d46627..cfd3279 100644
--- a/src/gallium/targets/xa-vmwgfx/Makefile
+++ b/src/gallium/targets/xa-vmwgfx/Makefile
@@ -4,7 +4,7 @@ include $(TOP)/configs/current
 ##### MACROS #####
 
 XA_MAJOR = 0
-XA_MINOR = 2
+XA_MINOR = 3
 XA_TINY = 0
 XA_CFLAGS = -g -fPIC
 
@@ -76,6 +76,7 @@ install: xatracker.pc
 	$(INSTALL) -d $(DESTDIR)$(INSTALL_LIB_DIR)/pkgconfig
 	$(INSTALL) -m 644 $(TOP)/src/gallium/state_trackers/xa/xa_tracker.h $(DESTDIR)$(INSTALL_INC_DIR)
 	$(INSTALL) -m 644 $(TOP)/src/gallium/state_trackers/xa/xa_context.h $(DESTDIR)$(INSTALL_INC_DIR)
+	$(INSTALL) -m 644 $(TOP)/src/gallium/state_trackers/xa/xa_composite.h $(DESTDIR)$(INSTALL_INC_DIR)
 	$(MINSTALL) -m 755 $(TOP)/$(LIB_DIR)/gallium/$(XA_LIB_GLOB) $(DESTDIR)$(INSTALL_LIB_DIR)
 	$(INSTALL) -m 644 xatracker.pc $(DESTDIR)$(INSTALL_LIB_DIR)/pkgconfig