st/xa: Render update. Better support for solid pictures

Remove unused and obsolete code for gradients and component-alpha
Support solid source- and mask pictures using a variable number
of samplers in the composite pipeline rather than the fixed number
we used before.

Tested using rendercheck for XA.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
This commit is contained in:
Thomas Hellstrom 2018-11-07 11:14:04 +01:00
parent 4bba280937
commit d86ad38205
5 changed files with 219 additions and 434 deletions

View File

@ -112,7 +112,7 @@ blend_for_op(struct xa_composite_blend *blend,
boolean supported = FALSE;
/*
* Temporarily disable component alpha since it appears buggy.
* No component alpha yet.
*/
if (mask_pic && mask_pic->component_alpha)
return FALSE;
@ -126,6 +126,7 @@ blend_for_op(struct xa_composite_blend *blend,
if (xa_blends[i].op == op) {
*blend = xa_blends[i];
supported = TRUE;
break;
}
}
@ -150,21 +151,6 @@ blend_for_op(struct xa_composite_blend *blend,
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(mask_pic->pict_format) &&
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;
}
@ -219,38 +205,26 @@ xa_composite_check_accelerated(const struct xa_composite *comp)
{
struct xa_composite_blend blend;
struct xa_picture *src_pic = comp->src;
struct xa_picture *mask_pic = comp->mask;
/*
* No component alpha yet.
*/
if (mask_pic && mask_pic->component_alpha)
return -XA_ERR_INVAL;
if (!xa_is_filter_accelerated(src_pic) ||
!xa_is_filter_accelerated(comp->mask)) {
return -XA_ERR_INVAL;
}
if (src_pic->src_pict &&src_pic->src_pict->type != xa_src_pict_solid_fill)
return -XA_ERR_INVAL;
if (src_pic->src_pict) {
if (src_pic->src_pict->type != xa_src_pict_solid_fill)
return -XA_ERR_INVAL;
if (!blend_for_op(&blend, comp->op, comp->src, comp->mask, comp->dst))
return -XA_ERR_INVAL;
/*
* Currently we don't support solid fill with a mask.
* We can easily do that, but that would require shader,
* sampler view setup and vertex setup modification.
*/
if (comp->mask)
return -XA_ERR_INVAL;
}
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(mask->pict_format)) {
if (blend.alpha_src && blend.rgb_src != PIPE_BLENDFACTOR_ZERO) {
return -XA_ERR_INVAL;
}
}
return XA_ERR_NONE;
}
return -XA_ERR_INVAL;
return XA_ERR_NONE;
}
static int
@ -293,7 +267,7 @@ picture_format_fixups(struct xa_picture *src_pic,
src_hw_format = xa_surface_format(src);
src_pic_format = src_pic->pict_format;
set_alpha = (xa_format_type_is_color(src_pic_format) &&
set_alpha = (xa_format_type_is_color(src_hw_format) &&
xa_format_a(src_pic_format) == 0);
if (set_alpha)
@ -324,6 +298,15 @@ picture_format_fixups(struct xa_picture *src_pic,
return ret;
}
static void
xa_src_in_mask(float src[4], const float mask[4])
{
src[0] *= mask[3];
src[1] *= mask[3];
src[2] *= mask[3];
src[3] *= mask[3];
}
static int
bind_shaders(struct xa_context *ctx, const struct xa_composite *comp)
{
@ -332,47 +315,56 @@ bind_shaders(struct xa_context *ctx, const struct xa_composite *comp)
struct xa_picture *src_pic = comp->src;
struct xa_picture *mask_pic = comp->mask;
ctx->has_solid_color = FALSE;
ctx->has_solid_src = FALSE;
ctx->has_solid_mask = FALSE;
if (src_pic) {
if (src_pic->wrap == xa_wrap_clamp_to_border && src_pic->has_transform)
fs_traits |= FS_SRC_REPEAT_NONE;
fs_traits |= FS_COMPOSITE;
vs_traits |= VS_COMPOSITE;
if (src_pic->src_pict) {
if (src_pic->src_pict->type == xa_src_pict_solid_fill) {
fs_traits |= FS_SOLID_FILL | FS_FILL;
vs_traits |= VS_SOLID_FILL;
fs_traits |= FS_SRC_SRC;
vs_traits |= VS_SRC_SRC;
xa_pixel_to_float4(src_pic->src_pict->solid_fill.color,
ctx->solid_color);
ctx->has_solid_color = TRUE;
ctx->has_solid_src = TRUE;
}
} else {
fs_traits |= FS_COMPOSITE;
vs_traits |= VS_COMPOSITE;
}
fs_traits |= picture_format_fixups(src_pic, 0);
} else
fs_traits |= picture_format_fixups(src_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->src_pict) {
if (mask_pic->src_pict->type == xa_src_pict_solid_fill) {
if (ctx->has_solid_src) {
float solid_mask[4];
if (mask_pic->component_alpha) {
struct xa_composite_blend blend;
if (!blend_for_op(&blend, comp->op, src_pic, mask_pic, NULL))
return -XA_ERR_INVAL;
xa_pixel_to_float4(mask_pic->src_pict->solid_fill.color,
solid_mask);
xa_src_in_mask(ctx->solid_color, solid_mask);
vs_traits &= ~(VS_MASK);
fs_traits &= ~(FS_MASK);
} else {
xa_pixel_to_float4(mask_pic->src_pict->solid_fill.color,
ctx->solid_color);
vs_traits |= VS_MASK_SRC;
fs_traits |= FS_MASK_SRC;
}
ctx->has_solid_mask = TRUE;
}
} else {
if (mask_pic->wrap == xa_wrap_clamp_to_border &&
mask_pic->has_transform)
fs_traits |= FS_MASK_REPEAT_NONE;
if (blend.alpha_src) {
fs_traits |= FS_CA_SRCALPHA;
} else
fs_traits |= FS_CA_FULL;
}
fs_traits |= picture_format_fixups(mask_pic, 1);
fs_traits |= picture_format_fixups(mask_pic, 1);
}
}
if (ctx->srf->format == PIPE_FORMAT_L8_UNORM ||
@ -396,42 +388,35 @@ bind_samplers(struct xa_context *ctx,
struct pipe_context *pipe = ctx->pipe;
struct xa_picture *src_pic = comp->src;
struct xa_picture *mask_pic = comp->mask;
int num_samplers = 0;
ctx->num_bound_samplers = 0;
xa_ctx_sampler_views_destroy(ctx);
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;
if (src_pic && !ctx->has_solid_src) {
unsigned src_wrap = xa_repeat_to_gallium(src_pic->wrap);
int filter;
(void) xa_filter_to_gallium(src_pic->filter, &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;
}
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;
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);
ctx->bound_sampler_views[0] = src_view;
num_samplers++;
}
if (mask_pic) {
unsigned mask_wrap = xa_repeat_to_gallium(mask_pic->wrap);
if (mask_pic && !ctx->has_solid_mask) {
unsigned mask_wrap = xa_repeat_to_gallium(mask_pic->wrap);
int filter;
(void) xa_filter_to_gallium(mask_pic->filter, &filter);
@ -442,31 +427,21 @@ bind_samplers(struct xa_context *ctx,
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;
samplers[num_samplers] = &mask_sampler;
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;
/*
* If src is a solid color, we have no src view, so set up a
* dummy one that will not be used anyway.
*/
if (ctx->bound_sampler_views[0] == NULL)
pipe_sampler_view_reference(&ctx->bound_sampler_views[0],
src_view);
ctx->bound_sampler_views[num_samplers] = src_view;
num_samplers++;
}
cso_set_samplers(ctx->cso, PIPE_SHADER_FRAGMENT, ctx->num_bound_samplers,
cso_set_samplers(ctx->cso, PIPE_SHADER_FRAGMENT, num_samplers,
(const struct pipe_sampler_state **)samplers);
cso_set_sampler_views(ctx->cso, PIPE_SHADER_FRAGMENT, ctx->num_bound_samplers,
cso_set_sampler_views(ctx->cso, PIPE_SHADER_FRAGMENT, num_samplers,
ctx->bound_sampler_views);
ctx->num_bound_samplers = num_samplers;
}
XA_EXPORT int
@ -476,9 +451,6 @@ xa_composite_prepare(struct xa_context *ctx,
struct xa_surface *dst_srf = comp->dst->srf;
int ret;
if (comp->mask && !comp->mask->srf)
return -XA_ERR_INVAL;
ret = xa_ctx_srf_create(ctx, dst_srf);
if (ret != XA_ERR_NONE)
return ret;
@ -511,8 +483,8 @@ xa_composite_rect(struct xa_context *ctx,
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);
xa_scissor_update(ctx, dstX, dstY, dstX + width, dstY + height);
renderer_solid(ctx, dstX, dstY, dstX + width, dstY + height);
} else {
const struct xa_composite *comp = ctx->comp;
int pos[6] = {srcX, srcY, maskX, maskY, dstX, dstY};
@ -537,7 +509,8 @@ xa_composite_done(struct xa_context *ctx)
renderer_draw_flush(ctx);
ctx->comp = NULL;
ctx->has_solid_color = FALSE;
ctx->has_solid_src = FALSE;
ctx->has_solid_mask = FALSE;
xa_ctx_sampler_views_destroy(ctx);
}

View File

@ -308,7 +308,7 @@ xa_solid_prepare(struct xa_context *ctx, struct xa_surface *dst,
xa_pixel_to_float4_a8(fg, ctx->solid_color);
else
xa_pixel_to_float4(fg, ctx->solid_color);
ctx->has_solid_color = 1;
ctx->has_solid_src = 1;
ctx->dst = dst;
@ -320,8 +320,8 @@ xa_solid_prepare(struct xa_context *ctx, struct xa_surface *dst,
exa->solid_color[2], exa->solid_color[3]);
#endif
vs_traits = VS_SOLID_FILL;
fs_traits = FS_SOLID_FILL;
vs_traits = VS_SRC_SRC | VS_COMPOSITE;
fs_traits = FS_SRC_SRC | VS_COMPOSITE;
renderer_bind_destination(ctx, ctx->srf);
bind_solid_blend_state(ctx);
@ -342,7 +342,7 @@ XA_EXPORT void
xa_solid(struct xa_context *ctx, int x, int y, int width, int height)
{
xa_scissor_update(ctx, x, y, x + width, y + height);
renderer_solid(ctx, x, y, x + width, y + height, ctx->solid_color);
renderer_solid(ctx, x, y, x + width, y + height);
}
XA_EXPORT void
@ -350,7 +350,7 @@ xa_solid_done(struct xa_context *ctx)
{
renderer_draw_flush(ctx);
ctx->comp = NULL;
ctx->has_solid_color = FALSE;
ctx->has_solid_src = FALSE;
ctx->num_bound_samplers = 0;
}

View File

@ -113,7 +113,8 @@ struct xa_context {
int simple_copy;
int has_solid_color;
int has_solid_src;
int has_solid_mask;
float solid_color[4];
unsigned int num_bound_samplers;
@ -145,35 +146,26 @@ xa_scissor_update(struct xa_context *ctx, unsigned minx, unsigned miny,
enum xa_vs_traits {
VS_COMPOSITE = 1 << 0,
VS_MASK = 1 << 1,
VS_SOLID_FILL = 1 << 2,
VS_LINGRAD_FILL = 1 << 3,
VS_RADGRAD_FILL = 1 << 4,
VS_YUV = 1 << 5,
VS_FILL = (VS_SOLID_FILL | VS_LINGRAD_FILL | VS_RADGRAD_FILL)
VS_SRC_SRC = 1 << 2,
VS_MASK_SRC = 1 << 3,
VS_YUV = 1 << 4,
};
enum xa_fs_traits {
FS_COMPOSITE = 1 << 0,
FS_MASK = 1 << 1,
FS_SOLID_FILL = 1 << 2,
FS_LINGRAD_FILL = 1 << 3,
FS_RADGRAD_FILL = 1 << 4,
FS_CA_FULL = 1 << 5, /* src.rgba * mask.rgba */
FS_CA_SRCALPHA = 1 << 6, /* src.aaaa * mask.rgba */
FS_YUV = 1 << 7,
FS_SRC_REPEAT_NONE = 1 << 8,
FS_MASK_REPEAT_NONE = 1 << 9,
FS_SRC_SWIZZLE_RGB = 1 << 10,
FS_MASK_SWIZZLE_RGB = 1 << 11,
FS_SRC_SET_ALPHA = 1 << 12,
FS_MASK_SET_ALPHA = 1 << 13,
FS_SRC_LUMINANCE = 1 << 14,
FS_MASK_LUMINANCE = 1 << 15,
FS_DST_LUMINANCE = 1 << 16,
FS_FILL = (FS_SOLID_FILL | FS_LINGRAD_FILL | FS_RADGRAD_FILL),
FS_COMPONENT_ALPHA = (FS_CA_FULL | FS_CA_SRCALPHA)
FS_SRC_SRC = 1 << 2,
FS_MASK_SRC = 1 << 3,
FS_YUV = 1 << 4,
FS_SRC_REPEAT_NONE = 1 << 5,
FS_MASK_REPEAT_NONE = 1 << 6,
FS_SRC_SWIZZLE_RGB = 1 << 7,
FS_MASK_SWIZZLE_RGB = 1 << 8,
FS_SRC_SET_ALPHA = 1 << 9,
FS_MASK_SET_ALPHA = 1 << 10,
FS_SRC_LUMINANCE = 1 << 11,
FS_MASK_LUMINANCE = 1 << 12,
FS_DST_LUMINANCE = 1 << 13,
};
struct xa_shader {
@ -282,7 +274,7 @@ 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);
int x0, int y0, int x1, int y1);
void
renderer_begin_textures(struct xa_context *r);

View File

@ -137,7 +137,7 @@ renderer_init_state(struct xa_context *r)
}
static inline void
add_vertex_color(struct xa_context *r, float x, float y, float color[4])
add_vertex_none(struct xa_context *r, float x, float y)
{
float *vertex = r->buffer + r->buffer_size;
@ -146,12 +146,7 @@ add_vertex_color(struct xa_context *r, float x, float y, float color[4])
vertex[2] = 0.f; /*z */
vertex[3] = 1.f; /*w */
vertex[4] = color[0]; /*r */
vertex[5] = color[1]; /*g */
vertex[6] = color[2]; /*b */
vertex[7] = color[3]; /*a */
r->buffer_size += 8;
r->buffer_size += 4;
}
static inline void
@ -554,27 +549,29 @@ void
renderer_begin_solid(struct xa_context *r)
{
r->buffer_size = 0;
r->attrs_per_vertex = 2;
r->attrs_per_vertex = 1;
renderer_set_constants(r, PIPE_SHADER_FRAGMENT, r->solid_color,
4 * sizeof(float));
}
void
renderer_solid(struct xa_context *r,
int x0, int y0, int x1, int y1, float *color)
int x0, int y0, int x1, int y1)
{
/*
* debug_printf("solid rect[(%d, %d), (%d, %d)], rgba[%f, %f, %f, %f]\n",
* x0, y0, x1, y1, color[0], color[1], color[2], color[3]); */
renderer_draw_conditional(r, 4 * 8);
renderer_draw_conditional(r, 4 * 4);
/* 1st vertex */
add_vertex_color(r, x0, y0, color);
add_vertex_none(r, x0, y0);
/* 2nd vertex */
add_vertex_color(r, x1, y0, color);
add_vertex_none(r, x1, y0);
/* 3rd vertex */
add_vertex_color(r, x1, y1, color);
add_vertex_none(r, x1, y1);
/* 4th vertex */
add_vertex_color(r, x0, y1, color);
add_vertex_none(r, x0, y1);
}
void
@ -588,6 +585,9 @@ renderer_begin_textures(struct xa_context *r)
{
r->attrs_per_vertex = 1 + r->num_bound_samplers;
r->buffer_size = 0;
if (r->has_solid_src || r->has_solid_mask)
renderer_set_constants(r, PIPE_SHADER_FRAGMENT, r->solid_color,
4 * sizeof(float));
}
void
@ -617,11 +617,19 @@ renderer_texture(struct xa_context *r,
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);
if (!r->has_solid_src) {
add_vertex_data1(r,
pos[0], pos[1], /* src */
pos[4], pos[5], /* dst */
width, height,
sampler_view[0]->texture, src_matrix);
} else {
add_vertex_data1(r,
pos[2], pos[3], /* mask */
pos[4], pos[5], /* dst */
width, height,
sampler_view[0]->texture, mask_matrix);
}
break;
case 3:
renderer_draw_conditional(r, 4 * 12);

View File

@ -48,19 +48,18 @@
* CONST[1] = (-1, -1, 0, 0)
*
* OUT[0] = vertex pos
* OUT[1] = src tex coord | solid fill color
* OUT[1] = src tex coord
* OUT[2] = mask tex coord
* OUT[3] = dst tex coord
*/
/* Fragment shader:
* SAMP[0] = src
* SAMP[1] = mask
* SAMP[2] = dst
* IN[0] = pos src | solid fill color
* IN[1] = pos mask
* IN[2] = pos dst
* CONST[0] = (0, 0, 0, 1)
/* Fragment shader. Samplers are allocated when needed.
* SAMP[0] = sampler for first texture (src or mask if src is solid)
* SAMP[1] = sampler for second texture (mask or none)
* IN[0] = first texture coordinates if present
* IN[1] = second texture coordinates if present
* CONST[0] = Solid color (src if src solid or mask if mask solid
* or src in mask if both solid).
*
* OUT[0] = color
*/
@ -71,21 +70,18 @@ print_fs_traits(int fs_traits)
const char *strings[] = {
"FS_COMPOSITE", /* = 1 << 0, */
"FS_MASK", /* = 1 << 1, */
"FS_SOLID_FILL", /* = 1 << 2, */
"FS_LINGRAD_FILL", /* = 1 << 3, */
"FS_RADGRAD_FILL", /* = 1 << 4, */
"FS_CA_FULL", /* = 1 << 5, *//* src.rgba * mask.rgba */
"FS_CA_SRCALPHA", /* = 1 << 6, *//* src.aaaa * mask.rgba */
"FS_YUV", /* = 1 << 7, */
"FS_SRC_REPEAT_NONE", /* = 1 << 8, */
"FS_MASK_REPEAT_NONE", /* = 1 << 9, */
"FS_SRC_SWIZZLE_RGB", /* = 1 << 10, */
"FS_MASK_SWIZZLE_RGB", /* = 1 << 11, */
"FS_SRC_SET_ALPHA", /* = 1 << 12, */
"FS_MASK_SET_ALPHA", /* = 1 << 13, */
"FS_SRC_LUMINANCE", /* = 1 << 14, */
"FS_MASK_LUMINANCE", /* = 1 << 15, */
"FS_DST_LUMINANCE", /* = 1 << 15, */
"FS_SRC_SRC", /* = 1 << 2, */
"FS_MASK_SRC", /* = 1 << 3, */
"FS_YUV", /* = 1 << 4, */
"FS_SRC_REPEAT_NONE", /* = 1 << 5, */
"FS_MASK_REPEAT_NONE", /* = 1 << 6, */
"FS_SRC_SWIZZLE_RGB", /* = 1 << 7, */
"FS_MASK_SWIZZLE_RGB", /* = 1 << 8, */
"FS_SRC_SET_ALPHA", /* = 1 << 9, */
"FS_MASK_SET_ALPHA", /* = 1 << 10, */
"FS_SRC_LUMINANCE", /* = 1 << 11, */
"FS_MASK_LUMINANCE", /* = 1 << 12, */
"FS_DST_LUMINANCE", /* = 1 << 13, */
};
int i, k;
@ -111,18 +107,12 @@ src_in_mask(struct ureg_program *ureg,
struct ureg_dst dst,
struct ureg_src src,
struct ureg_src mask,
unsigned component_alpha, unsigned mask_luminance)
unsigned mask_luminance)
{
if (component_alpha == FS_CA_FULL) {
ureg_MUL(ureg, dst, src, mask);
} else if (component_alpha == FS_CA_SRCALPHA) {
ureg_MUL(ureg, dst, ureg_scalar(src, TGSI_SWIZZLE_W), mask);
} else {
if (mask_luminance)
ureg_MUL(ureg, dst, src, ureg_scalar(mask, TGSI_SWIZZLE_X));
else
ureg_MUL(ureg, dst, src, ureg_scalar(mask, TGSI_SWIZZLE_W));
}
if (mask_luminance)
ureg_MUL(ureg, dst, src, ureg_scalar(mask, TGSI_SWIZZLE_X));
else
ureg_MUL(ureg, dst, src, ureg_scalar(mask, TGSI_SWIZZLE_W));
}
static struct ureg_src
@ -139,125 +129,6 @@ vs_normalize_coords(struct ureg_program *ureg,
return ret;
}
static void
linear_gradient(struct ureg_program *ureg,
struct ureg_dst out,
struct ureg_src pos,
struct ureg_src sampler,
struct ureg_src coords,
struct ureg_src const0124,
struct ureg_src matrow0,
struct ureg_src matrow1, struct ureg_src matrow2)
{
struct ureg_dst temp0 = ureg_DECL_temporary(ureg);
struct ureg_dst temp1 = ureg_DECL_temporary(ureg);
struct ureg_dst temp2 = ureg_DECL_temporary(ureg);
struct ureg_dst temp3 = ureg_DECL_temporary(ureg);
struct ureg_dst temp4 = ureg_DECL_temporary(ureg);
struct ureg_dst temp5 = ureg_DECL_temporary(ureg);
ureg_MOV(ureg, ureg_writemask(temp0, TGSI_WRITEMASK_XY), pos);
ureg_MOV(ureg,
ureg_writemask(temp0, TGSI_WRITEMASK_Z),
ureg_scalar(const0124, TGSI_SWIZZLE_Y));
ureg_DP3(ureg, temp1, matrow0, ureg_src(temp0));
ureg_DP3(ureg, temp2, matrow1, ureg_src(temp0));
ureg_DP3(ureg, temp3, matrow2, ureg_src(temp0));
ureg_RCP(ureg, temp3, ureg_src(temp3));
ureg_MUL(ureg, temp1, ureg_src(temp1), ureg_src(temp3));
ureg_MUL(ureg, temp2, ureg_src(temp2), ureg_src(temp3));
ureg_MOV(ureg, ureg_writemask(temp4, TGSI_WRITEMASK_X), ureg_src(temp1));
ureg_MOV(ureg, ureg_writemask(temp4, TGSI_WRITEMASK_Y), ureg_src(temp2));
ureg_MUL(ureg, temp0,
ureg_scalar(coords, TGSI_SWIZZLE_Y),
ureg_scalar(ureg_src(temp4), TGSI_SWIZZLE_Y));
ureg_MAD(ureg, temp1,
ureg_scalar(coords, TGSI_SWIZZLE_X),
ureg_scalar(ureg_src(temp4), TGSI_SWIZZLE_X), ureg_src(temp0));
ureg_MUL(ureg, temp2, ureg_src(temp1), ureg_scalar(coords, TGSI_SWIZZLE_Z));
ureg_TEX(ureg, out, TGSI_TEXTURE_1D, ureg_src(temp2), sampler);
ureg_release_temporary(ureg, temp0);
ureg_release_temporary(ureg, temp1);
ureg_release_temporary(ureg, temp2);
ureg_release_temporary(ureg, temp3);
ureg_release_temporary(ureg, temp4);
ureg_release_temporary(ureg, temp5);
}
static void
radial_gradient(struct ureg_program *ureg,
struct ureg_dst out,
struct ureg_src pos,
struct ureg_src sampler,
struct ureg_src coords,
struct ureg_src const0124,
struct ureg_src matrow0,
struct ureg_src matrow1, struct ureg_src matrow2)
{
struct ureg_dst temp0 = ureg_DECL_temporary(ureg);
struct ureg_dst temp1 = ureg_DECL_temporary(ureg);
struct ureg_dst temp2 = ureg_DECL_temporary(ureg);
struct ureg_dst temp3 = ureg_DECL_temporary(ureg);
struct ureg_dst temp4 = ureg_DECL_temporary(ureg);
struct ureg_dst temp5 = ureg_DECL_temporary(ureg);
ureg_MOV(ureg, ureg_writemask(temp0, TGSI_WRITEMASK_XY), pos);
ureg_MOV(ureg,
ureg_writemask(temp0, TGSI_WRITEMASK_Z),
ureg_scalar(const0124, TGSI_SWIZZLE_Y));
ureg_DP3(ureg, temp1, matrow0, ureg_src(temp0));
ureg_DP3(ureg, temp2, matrow1, ureg_src(temp0));
ureg_DP3(ureg, temp3, matrow2, ureg_src(temp0));
ureg_RCP(ureg, temp3, ureg_src(temp3));
ureg_MUL(ureg, temp1, ureg_src(temp1), ureg_src(temp3));
ureg_MUL(ureg, temp2, ureg_src(temp2), ureg_src(temp3));
ureg_MOV(ureg, ureg_writemask(temp5, TGSI_WRITEMASK_X), ureg_src(temp1));
ureg_MOV(ureg, ureg_writemask(temp5, TGSI_WRITEMASK_Y), ureg_src(temp2));
ureg_MUL(ureg, temp0, ureg_scalar(coords, TGSI_SWIZZLE_Y),
ureg_scalar(ureg_src(temp5), TGSI_SWIZZLE_Y));
ureg_MAD(ureg, temp1,
ureg_scalar(coords, TGSI_SWIZZLE_X),
ureg_scalar(ureg_src(temp5), TGSI_SWIZZLE_X), ureg_src(temp0));
ureg_ADD(ureg, temp1, ureg_src(temp1), ureg_src(temp1));
ureg_MUL(ureg, temp3,
ureg_scalar(ureg_src(temp5), TGSI_SWIZZLE_Y),
ureg_scalar(ureg_src(temp5), TGSI_SWIZZLE_Y));
ureg_MAD(ureg, temp4,
ureg_scalar(ureg_src(temp5), TGSI_SWIZZLE_X),
ureg_scalar(ureg_src(temp5), TGSI_SWIZZLE_X), ureg_src(temp3));
ureg_MOV(ureg, temp4, ureg_negate(ureg_src(temp4)));
ureg_MUL(ureg, temp2, ureg_scalar(coords, TGSI_SWIZZLE_Z), ureg_src(temp4));
ureg_MUL(ureg, temp0,
ureg_scalar(const0124, TGSI_SWIZZLE_W), ureg_src(temp2));
ureg_MUL(ureg, temp3, ureg_src(temp1), ureg_src(temp1));
ureg_ADD(ureg, temp2, ureg_src(temp3), ureg_negate(ureg_src(temp0)));
ureg_RSQ(ureg, temp2, ureg_abs(ureg_src(temp2)));
ureg_RCP(ureg, temp2, ureg_src(temp2));
ureg_ADD(ureg, temp1, ureg_src(temp2), ureg_negate(ureg_src(temp1)));
ureg_ADD(ureg, temp0,
ureg_scalar(coords, TGSI_SWIZZLE_Z),
ureg_scalar(coords, TGSI_SWIZZLE_Z));
ureg_RCP(ureg, temp0, ureg_src(temp0));
ureg_MUL(ureg, temp2, ureg_src(temp1), ureg_src(temp0));
ureg_TEX(ureg, out, TGSI_TEXTURE_1D, ureg_src(temp2), sampler);
ureg_release_temporary(ureg, temp0);
ureg_release_temporary(ureg, temp1);
ureg_release_temporary(ureg, temp2);
ureg_release_temporary(ureg, temp3);
ureg_release_temporary(ureg, temp4);
ureg_release_temporary(ureg, temp5);
}
static void *
create_vs(struct pipe_context *pipe, unsigned vs_traits)
{
@ -265,10 +136,11 @@ create_vs(struct pipe_context *pipe, unsigned vs_traits)
struct ureg_src src;
struct ureg_dst dst;
struct ureg_src const0, const1;
boolean is_fill = (vs_traits & VS_FILL) != 0;
boolean is_composite = (vs_traits & VS_COMPOSITE) != 0;
boolean has_mask = (vs_traits & VS_MASK) != 0;
boolean is_yuv = (vs_traits & VS_YUV) != 0;
boolean is_src_src = (vs_traits & VS_SRC_SRC) != 0;
boolean is_mask_src = (vs_traits & VS_MASK_SRC) != 0;
unsigned input_slot = 0;
ureg = ureg_create(PIPE_SHADER_VERTEX);
@ -279,8 +151,6 @@ create_vs(struct pipe_context *pipe, unsigned vs_traits)
const1 = ureg_DECL_constant(ureg, 1);
/* it has to be either a fill or a composite op */
debug_assert((is_fill ^ is_composite) ^ is_yuv);
src = ureg_DECL_vs_input(ureg, input_slot++);
dst = ureg_DECL_output(ureg, TGSI_SEMANTIC_POSITION, 0);
src = vs_normalize_coords(ureg, src, const0, const1);
@ -293,21 +163,17 @@ create_vs(struct pipe_context *pipe, unsigned vs_traits)
}
if (is_composite) {
src = ureg_DECL_vs_input(ureg, input_slot++);
dst = ureg_DECL_output(ureg, TGSI_SEMANTIC_GENERIC, 0);
ureg_MOV(ureg, dst, src);
}
if (!is_src_src || (has_mask && !is_mask_src)) {
src = ureg_DECL_vs_input(ureg, input_slot++);
dst = ureg_DECL_output(ureg, TGSI_SEMANTIC_GENERIC, 0);
ureg_MOV(ureg, dst, src);
}
if (is_fill) {
src = ureg_DECL_vs_input(ureg, input_slot++);
dst = ureg_DECL_output(ureg, TGSI_SEMANTIC_COLOR, 0);
ureg_MOV(ureg, dst, src);
}
if (has_mask) {
src = ureg_DECL_vs_input(ureg, input_slot++);
dst = ureg_DECL_output(ureg, TGSI_SEMANTIC_GENERIC, 1);
ureg_MOV(ureg, dst, src);
if (!is_src_src && (has_mask && !is_mask_src)) {
src = ureg_DECL_vs_input(ureg, input_slot++);
dst = ureg_DECL_output(ureg, TGSI_SEMANTIC_GENERIC, 1);
ureg_MOV(ureg, dst, src);
}
}
ureg_END(ureg);
@ -383,7 +249,7 @@ xrender_tex(struct ureg_program *ureg,
struct ureg_dst dst,
struct ureg_src coords,
struct ureg_src sampler,
struct ureg_src imm0,
const struct ureg_src *imm0,
boolean repeat_none, boolean swizzle, boolean set_alpha)
{
if (repeat_none) {
@ -394,11 +260,11 @@ xrender_tex(struct ureg_program *ureg,
TGSI_SWIZZLE_X,
TGSI_SWIZZLE_Y,
TGSI_SWIZZLE_X,
TGSI_SWIZZLE_Y), ureg_scalar(imm0,
TGSI_SWIZZLE_Y), ureg_scalar(*imm0,
TGSI_SWIZZLE_X));
ureg_SLT(ureg, tmp0,
ureg_swizzle(coords, TGSI_SWIZZLE_X, TGSI_SWIZZLE_Y,
TGSI_SWIZZLE_X, TGSI_SWIZZLE_Y), ureg_scalar(imm0,
TGSI_SWIZZLE_X, TGSI_SWIZZLE_Y), ureg_scalar(*imm0,
TGSI_SWIZZLE_W));
ureg_MIN(ureg, tmp0, ureg_src(tmp0), ureg_src(tmp1));
ureg_MIN(ureg, tmp0, ureg_scalar(ureg_src(tmp0), TGSI_SWIZZLE_X),
@ -412,7 +278,7 @@ xrender_tex(struct ureg_program *ureg,
if (set_alpha)
ureg_MOV(ureg,
ureg_writemask(tmp1, TGSI_WRITEMASK_W),
ureg_scalar(imm0, TGSI_SWIZZLE_W));
ureg_scalar(*imm0, TGSI_SWIZZLE_W));
ureg_MUL(ureg, dst, ureg_src(tmp1), ureg_src(tmp0));
ureg_release_temporary(ureg, tmp0);
ureg_release_temporary(ureg, tmp1);
@ -432,7 +298,32 @@ xrender_tex(struct ureg_program *ureg,
if (set_alpha)
ureg_MOV(ureg,
ureg_writemask(dst, TGSI_WRITEMASK_W),
ureg_scalar(imm0, TGSI_SWIZZLE_W));
ureg_scalar(*imm0, TGSI_SWIZZLE_W));
}
}
static void
read_input(struct ureg_program *ureg,
struct ureg_dst dst,
const struct ureg_src *imm0,
boolean repeat_none, boolean swizzle, boolean set_alpha,
boolean is_src, unsigned *cur_constant, unsigned *cur_sampler)
{
struct ureg_src input, sampler;
if (is_src) {
input = ureg_DECL_constant(ureg, (*cur_constant)++);
ureg_MOV(ureg, dst, input);
} else {
sampler = ureg_DECL_sampler(ureg, *cur_sampler);
ureg_DECL_sampler_view(ureg, *cur_sampler, TGSI_TEXTURE_2D,
TGSI_RETURN_TYPE_FLOAT, TGSI_RETURN_TYPE_FLOAT,
TGSI_RETURN_TYPE_FLOAT, TGSI_RETURN_TYPE_FLOAT);
input = ureg_DECL_fs_input(ureg,
TGSI_SEMANTIC_GENERIC, (*cur_sampler)++,
TGSI_INTERPOLATE_PERSPECTIVE);
xrender_tex(ureg, dst, input, sampler, imm0,
repeat_none, swizzle, set_alpha);
}
}
@ -440,18 +331,10 @@ static void *
create_fs(struct pipe_context *pipe, unsigned fs_traits)
{
struct ureg_program *ureg;
struct ureg_src /*dst_sampler, */ src_sampler, mask_sampler;
struct ureg_src /*dst_pos, */ src_input, mask_pos;
struct ureg_dst src, mask;
struct ureg_dst out;
struct ureg_src imm0 = { 0 };
unsigned has_mask = (fs_traits & FS_MASK) != 0;
unsigned is_fill = (fs_traits & FS_FILL) != 0;
unsigned is_composite = (fs_traits & FS_COMPOSITE) != 0;
unsigned is_solid = (fs_traits & FS_SOLID_FILL) != 0;
unsigned is_lingrad = (fs_traits & FS_LINGRAD_FILL) != 0;
unsigned is_radgrad = (fs_traits & FS_RADGRAD_FILL) != 0;
unsigned comp_alpha_mask = fs_traits & FS_COMPONENT_ALPHA;
unsigned is_yuv = (fs_traits & FS_YUV) != 0;
unsigned src_repeat_none = (fs_traits & FS_SRC_REPEAT_NONE) != 0;
unsigned mask_repeat_none = (fs_traits & FS_MASK_REPEAT_NONE) != 0;
@ -462,6 +345,10 @@ create_fs(struct pipe_context *pipe, unsigned fs_traits)
unsigned src_luminance = (fs_traits & FS_SRC_LUMINANCE) != 0;
unsigned mask_luminance = (fs_traits & FS_MASK_LUMINANCE) != 0;
unsigned dst_luminance = (fs_traits & FS_DST_LUMINANCE) != 0;
unsigned is_src_src = (fs_traits & FS_SRC_SRC) != 0;
unsigned is_mask_src = (fs_traits & FS_MASK_SRC) != 0;
unsigned cur_sampler = 0;
unsigned cur_constant = 0;
#if 0
print_fs_traits(fs_traits);
@ -473,9 +360,8 @@ create_fs(struct pipe_context *pipe, unsigned fs_traits)
if (ureg == NULL)
return 0;
/* it has to be either a fill, a composite op or a yuv conversion */
debug_assert((is_fill ^ is_composite) ^ is_yuv);
(void)is_yuv;
if (is_yuv)
return create_yuv_shader(pipe, ureg);
out = ureg_DECL_output(ureg, TGSI_SEMANTIC_COLOR, 0);
@ -483,86 +369,13 @@ create_fs(struct pipe_context *pipe, unsigned fs_traits)
src_set_alpha || mask_set_alpha || src_luminance) {
imm0 = ureg_imm4f(ureg, 0, 0, 0, 1);
}
if (is_composite) {
src_sampler = ureg_DECL_sampler(ureg, 0);
ureg_DECL_sampler_view(ureg, 0, TGSI_TEXTURE_2D,
TGSI_RETURN_TYPE_FLOAT, TGSI_RETURN_TYPE_FLOAT,
TGSI_RETURN_TYPE_FLOAT, TGSI_RETURN_TYPE_FLOAT);
src_input = ureg_DECL_fs_input(ureg,
TGSI_SEMANTIC_GENERIC, 0,
TGSI_INTERPOLATE_PERSPECTIVE);
} else if (is_fill) {
if (is_solid)
src_input = ureg_DECL_fs_input(ureg,
TGSI_SEMANTIC_COLOR, 0,
TGSI_INTERPOLATE_PERSPECTIVE);
else
src_input = ureg_DECL_fs_input(ureg,
TGSI_SEMANTIC_POSITION, 0,
TGSI_INTERPOLATE_PERSPECTIVE);
} else {
debug_assert(is_yuv);
return create_yuv_shader(pipe, ureg);
}
if (has_mask) {
mask_sampler = ureg_DECL_sampler(ureg, 1);
ureg_DECL_sampler_view(ureg, 1, TGSI_TEXTURE_2D,
TGSI_RETURN_TYPE_FLOAT, TGSI_RETURN_TYPE_FLOAT,
TGSI_RETURN_TYPE_FLOAT, TGSI_RETURN_TYPE_FLOAT);
mask_pos = ureg_DECL_fs_input(ureg,
TGSI_SEMANTIC_GENERIC, 1,
TGSI_INTERPOLATE_PERSPECTIVE);
}
#if 0 /* unused right now */
dst_sampler = ureg_DECL_sampler(ureg, 2);
ureg_DECL_sampler_view(ureg, 2, TGSI_TEXTURE_2D,
TGSI_RETURN_TYPE_FLOAT, TGSI_RETURN_TYPE_FLOAT,
TGSI_RETURN_TYPE_FLOAT, TGSI_RETURN_TYPE_FLOAT);
dst_pos = ureg_DECL_fs_input(ureg,
TGSI_SEMANTIC_POSITION, 2,
TGSI_INTERPOLATE_PERSPECTIVE);
#endif
src = (has_mask || src_luminance || dst_luminance) ?
ureg_DECL_temporary(ureg) : out;
if (is_composite) {
if (has_mask || src_luminance || dst_luminance)
src = ureg_DECL_temporary(ureg);
else
src = out;
xrender_tex(ureg, src, src_input, src_sampler, imm0,
src_repeat_none, src_swizzle, src_set_alpha);
} else if (is_fill) {
if (is_solid) {
if (has_mask || src_luminance || dst_luminance)
src = ureg_dst(src_input);
else
ureg_MOV(ureg, out, src_input);
} else if (is_lingrad || is_radgrad) {
struct ureg_src coords, const0124, matrow0, matrow1, matrow2;
read_input(ureg, src, &imm0, src_repeat_none, src_swizzle,
src_set_alpha, is_src_src, &cur_constant, &cur_sampler);
if (has_mask || src_luminance || dst_luminance)
src = ureg_DECL_temporary(ureg);
else
src = out;
coords = ureg_DECL_constant(ureg, 0);
const0124 = ureg_DECL_constant(ureg, 1);
matrow0 = ureg_DECL_constant(ureg, 2);
matrow1 = ureg_DECL_constant(ureg, 3);
matrow2 = ureg_DECL_constant(ureg, 4);
if (is_lingrad) {
linear_gradient(ureg, src,
src_input, src_sampler,
coords, const0124, matrow0, matrow1, matrow2);
} else if (is_radgrad) {
radial_gradient(ureg, src,
src_input, src_sampler,
coords, const0124, matrow0, matrow1, matrow2);
}
} else
debug_assert(!"Unknown fill type!");
}
if (src_luminance) {
ureg_MOV(ureg, src, ureg_scalar(ureg_src(src), TGSI_SWIZZLE_X));
ureg_MOV(ureg, ureg_writemask(src, TGSI_WRITEMASK_XYZ),
@ -573,13 +386,12 @@ create_fs(struct pipe_context *pipe, unsigned fs_traits)
if (has_mask) {
mask = ureg_DECL_temporary(ureg);
xrender_tex(ureg, mask, mask_pos, mask_sampler, imm0,
mask_repeat_none, mask_swizzle, mask_set_alpha);
/* src IN mask */
read_input(ureg, mask, &imm0, mask_repeat_none,
mask_swizzle, mask_set_alpha, is_mask_src, &cur_constant,
&cur_sampler);
src_in_mask(ureg, (dst_luminance) ? src : out, ureg_src(src),
ureg_src(mask),
comp_alpha_mask, mask_luminance);
ureg_src(mask), mask_luminance);
ureg_release_temporary(ureg, mask);
}