/**
* Create a TGSI ureg_dst register from a Mesa dest register.
*/
static struct ureg_dst
translate_dst( struct st_translate *t,
const struct prog_dst_register *DstReg,
boolean saturate,
boolean clamp_color)
{
struct ureg_dst dst = dst_register( t,
DstReg->File,
DstReg->Index );
dst = ureg_writemask( dst,
DstReg->WriteMask );
if (saturate)
dst = ureg_saturate( dst );
else if (clamp_color && DstReg->File == PROGRAM_OUTPUT) {
/* Clamp colors for ARB_color_buffer_float. */
switch (t->procType) {
case TGSI_PROCESSOR_VERTEX:
/* This can only occur with a compatibility profile, which doesn't
* support geometry shaders. */
if (DstReg->Index == VARYING_SLOT_COL0 ||
DstReg->Index == VARYING_SLOT_COL1 ||
DstReg->Index == VARYING_SLOT_BFC0 ||
DstReg->Index == VARYING_SLOT_BFC1) {
dst = ureg_saturate(dst);
}
break;
case TGSI_PROCESSOR_FRAGMENT:
if (DstReg->Index >= FRAG_RESULT_COLOR) {
dst = ureg_saturate(dst);
}
break;
}
}
if (DstReg->RelAddr)
dst = ureg_dst_indirect( dst, ureg_src(t->address[0]) );
return dst;
}
/**
* Create a TGSI ureg_dst register from a Mesa dest register.
*/
static struct ureg_dst
translate_dst( struct st_translate *t,
const struct prog_dst_register *DstReg,
boolean saturate,
boolean clamp_color)
{
struct ureg_dst dst = dst_register( t,
DstReg->File,
DstReg->Index );
dst = ureg_writemask( dst,
DstReg->WriteMask );
if (saturate)
dst = ureg_saturate( dst );
else if (clamp_color && DstReg->File == PROGRAM_OUTPUT) {
/* Clamp colors for ARB_color_buffer_float. */
switch (t->procType) {
case TGSI_PROCESSOR_VERTEX:
/* XXX if the geometry shader is present, this must be done there
* instead of here. */
if (DstReg->Index == VARYING_SLOT_COL0 ||
DstReg->Index == VARYING_SLOT_COL1 ||
DstReg->Index == VARYING_SLOT_BFC0 ||
DstReg->Index == VARYING_SLOT_BFC1) {
dst = ureg_saturate(dst);
}
break;
case TGSI_PROCESSOR_FRAGMENT:
if (DstReg->Index >= FRAG_RESULT_COLOR) {
dst = ureg_saturate(dst);
}
break;
}
}
if (DstReg->RelAddr)
dst = ureg_dst_indirect( dst, ureg_src(t->address[0]) );
return dst;
}
/**
* OpenGL's fragment gl_FrontFace input is 1 for front-facing, 0 for back.
* TGSI uses +1 for front, -1 for back.
* This function converts the TGSI value to the GL value. Simply clamping/
* saturating the value to [0,1] does the job.
*/
static void
emit_face_var( struct st_translate *t,
const struct gl_program *program )
{
struct ureg_program *ureg = t->ureg;
struct ureg_dst face_temp = ureg_DECL_temporary( ureg );
struct ureg_src face_input = t->inputs[t->inputMapping[VARYING_SLOT_FACE]];
/* MOV_SAT face_temp, input[face]
*/
face_temp = ureg_saturate( face_temp );
ureg_MOV( ureg, face_temp, face_input );
/* Use face_temp as face input from here on:
*/
t->inputs[t->inputMapping[VARYING_SLOT_FACE]] = ureg_src(face_temp);
}
static void
emit_dstmod(struct st_translate *t,
struct ureg_dst dst, GLuint dstMod)
{
float imm;
struct ureg_src src[3];
GLuint scale = dstMod & ~GL_SATURATE_BIT_ATI;
if (dstMod == GL_NONE) {
return;
}
switch (scale) {
case GL_2X_BIT_ATI:
imm = 2.0f;
break;
case GL_4X_BIT_ATI:
imm = 4.0f;
break;
case GL_8X_BIT_ATI:
imm = 8.0f;
break;
case GL_HALF_BIT_ATI:
imm = 0.5f;
break;
case GL_QUARTER_BIT_ATI:
imm = 0.25f;
break;
case GL_EIGHTH_BIT_ATI:
imm = 0.125f;
break;
default:
imm = 1.0f;
}
src[0] = ureg_src(dst);
src[1] = ureg_imm1f(t->ureg, imm);
if (dstMod & GL_SATURATE_BIT_ATI) {
dst = ureg_saturate(dst);
}
ureg_insn(t->ureg, TGSI_OPCODE_MUL, &dst, 1, src, 2);
}
/**
* Create a TGSI ureg_dst register from a Mesa dest register.
*/
static struct ureg_dst
translate_dst( struct st_translate *t,
const struct prog_dst_register *DstReg,
boolean saturate)
{
struct ureg_dst dst = dst_register( t,
DstReg->File,
DstReg->Index );
dst = ureg_writemask( dst,
DstReg->WriteMask );
if (saturate)
dst = ureg_saturate( dst );
if (DstReg->RelAddr)
dst = ureg_dst_indirect( dst, ureg_src(t->address[0]) );
return dst;
}
static void *
create_deint_frag_shader(struct vl_deint_filter *filter, unsigned field,
struct vertex2f *sizes, bool spatial_filter)
{
struct ureg_program *shader;
struct ureg_src i_vtex;
struct ureg_src sampler_cur;
struct ureg_src sampler_prevprev;
struct ureg_src sampler_prev;
struct ureg_src sampler_next;
struct ureg_dst o_fragment;
struct ureg_dst t_tex;
struct ureg_dst t_comp_top, t_comp_bot;
struct ureg_dst t_diff;
struct ureg_dst t_a, t_b;
struct ureg_dst t_weave, t_linear;
shader = ureg_create(PIPE_SHADER_FRAGMENT);
if (!shader) {
return NULL;
}
t_tex = ureg_DECL_temporary(shader);
t_comp_top = ureg_DECL_temporary(shader);
t_comp_bot = ureg_DECL_temporary(shader);
t_diff = ureg_DECL_temporary(shader);
t_a = ureg_DECL_temporary(shader);
t_b = ureg_DECL_temporary(shader);
t_weave = ureg_DECL_temporary(shader);
t_linear = ureg_DECL_temporary(shader);
i_vtex = ureg_DECL_fs_input(shader, TGSI_SEMANTIC_GENERIC, VS_O_VTEX, TGSI_INTERPOLATE_LINEAR);
sampler_prevprev = ureg_DECL_sampler(shader, 0);
sampler_prev = ureg_DECL_sampler(shader, 1);
sampler_cur = ureg_DECL_sampler(shader, 2);
sampler_next = ureg_DECL_sampler(shader, 3);
o_fragment = ureg_DECL_output(shader, TGSI_SEMANTIC_COLOR, 0);
// we don't care about ZW interpolation (allows better optimization)
ureg_MOV(shader, t_tex, i_vtex);
ureg_MOV(shader, ureg_writemask(t_tex, TGSI_WRITEMASK_ZW),
ureg_imm1f(shader, 0));
// sample between texels for cheap lowpass
ureg_ADD(shader, t_comp_top, ureg_src(t_tex),
ureg_imm4f(shader, sizes->x * 0.5f, sizes->y * -0.5f, 0, 0));
ureg_ADD(shader, t_comp_bot, ureg_src(t_tex),
ureg_imm4f(shader, sizes->x * -0.5f, sizes->y * 0.5f, 1.0f, 0));
if (field == 0) {
/* interpolating top field -> current field is a bottom field */
// cur vs prev2
ureg_TEX(shader, t_a, TGSI_TEXTURE_2D_ARRAY, ureg_src(t_comp_bot), sampler_cur);
ureg_TEX(shader, t_b, TGSI_TEXTURE_2D_ARRAY, ureg_src(t_comp_bot), sampler_prevprev);
ureg_ADD(shader, ureg_writemask(t_diff, TGSI_WRITEMASK_X), ureg_src(t_a), ureg_negate(ureg_src(t_b)));
// prev vs next
ureg_TEX(shader, t_a, TGSI_TEXTURE_2D_ARRAY, ureg_src(t_comp_top), sampler_prev);
ureg_TEX(shader, t_b, TGSI_TEXTURE_2D_ARRAY, ureg_src(t_comp_top), sampler_next);
ureg_ADD(shader, ureg_writemask(t_diff, TGSI_WRITEMASK_Y), ureg_src(t_a), ureg_negate(ureg_src(t_b)));
} else {
/* interpolating bottom field -> current field is a top field */
// cur vs prev2
ureg_TEX(shader, t_a, TGSI_TEXTURE_2D_ARRAY, ureg_src(t_comp_top), sampler_cur);
ureg_TEX(shader, t_b, TGSI_TEXTURE_2D_ARRAY, ureg_src(t_comp_top), sampler_prevprev);
ureg_ADD(shader, ureg_writemask(t_diff, TGSI_WRITEMASK_X), ureg_src(t_a), ureg_negate(ureg_src(t_b)));
// prev vs next
ureg_TEX(shader, t_a, TGSI_TEXTURE_2D_ARRAY, ureg_src(t_comp_bot), sampler_prev);
ureg_TEX(shader, t_b, TGSI_TEXTURE_2D_ARRAY, ureg_src(t_comp_bot), sampler_next);
ureg_ADD(shader, ureg_writemask(t_diff, TGSI_WRITEMASK_Y), ureg_src(t_a), ureg_negate(ureg_src(t_b)));
}
// absolute maximum of differences
ureg_MAX(shader, ureg_writemask(t_diff, TGSI_WRITEMASK_X), ureg_abs(ureg_src(t_diff)),
ureg_scalar(ureg_abs(ureg_src(t_diff)), TGSI_SWIZZLE_Y));
if (field == 0) {
/* weave with prev top field */
ureg_TEX(shader, t_weave, TGSI_TEXTURE_2D_ARRAY, ureg_src(t_tex), sampler_prev);
/* get linear interpolation from current bottom field */
ureg_ADD(shader, t_comp_top, ureg_src(t_tex), ureg_imm4f(shader, 0, sizes->y * -1.0f, 1.0f, 0));
ureg_TEX(shader, t_linear, TGSI_TEXTURE_2D_ARRAY, ureg_src(t_comp_top), sampler_cur);
} else {
/* weave with prev bottom field */
ureg_ADD(shader, t_comp_bot, ureg_src(t_tex), ureg_imm4f(shader, 0, 0, 1.0f, 0));
ureg_TEX(shader, t_weave, TGSI_TEXTURE_2D_ARRAY, ureg_src(t_comp_bot), sampler_prev);
/* get linear interpolation from current top field */
ureg_ADD(shader, t_comp_bot, ureg_src(t_tex), ureg_imm4f(shader, 0, sizes->y * 1.0f, 0, 0));
ureg_TEX(shader, t_linear, TGSI_TEXTURE_2D_ARRAY, ureg_src(t_comp_bot), sampler_cur);
}
// mix between weave and linear
// fully weave if diff < 6 (0.02353), fully interpolate if diff > 14 (0.05490)
ureg_ADD(shader, ureg_writemask(t_diff, TGSI_WRITEMASK_X), ureg_src(t_diff),
ureg_imm4f(shader, -0.02353f, 0, 0, 0));
ureg_MUL(shader, ureg_saturate(ureg_writemask(t_diff, TGSI_WRITEMASK_X)),
ureg_src(t_diff), ureg_imm4f(shader, 31.8750f, 0, 0, 0));
ureg_LRP(shader, ureg_writemask(t_tex, TGSI_WRITEMASK_X), ureg_src(t_diff),
ureg_src(t_linear), ureg_src(t_weave));
ureg_MOV(shader, o_fragment, ureg_scalar(ureg_src(t_tex), TGSI_SWIZZLE_X));
ureg_release_temporary(shader, t_tex);
ureg_release_temporary(shader, t_comp_top);
ureg_release_temporary(shader, t_comp_bot);
ureg_release_temporary(shader, t_diff);
ureg_release_temporary(shader, t_a);
ureg_release_temporary(shader, t_b);
ureg_release_temporary(shader, t_weave);
ureg_release_temporary(shader, t_linear);
ureg_END(shader);
return ureg_create_shader_and_destroy(shader, filter->pipe);
}