static void build_fog( struct tnl_program *p )
{
struct ureg fog = register_output(p, VERT_RESULT_FOGC);
struct ureg input;
GLuint useabs = p->state->fog_source_is_depth && p->state->fog_option &&
(p->state->fog_option != FOG_EXP2);
if (p->state->fog_source_is_depth) {
input = swizzle1(get_eye_position(p), Z);
}
else {
input = swizzle1(register_input(p, VERT_ATTRIB_FOG), X);
}
if (p->state->fog_option &&
p->state->tnl_do_vertex_fog) {
struct ureg params = register_param2(p, STATE_INTERNAL,
STATE_FOG_PARAMS_OPTIMIZED);
struct ureg tmp = get_temp(p);
struct ureg id = get_identity_param(p);
emit_op1(p, OPCODE_MOV, fog, 0, id);
if (useabs) {
emit_op1(p, OPCODE_ABS, tmp, 0, input);
}
switch (p->state->fog_option) {
case FOG_LINEAR: {
emit_op3(p, OPCODE_MAD, tmp, 0, useabs ? tmp : input,
swizzle1(params,X), swizzle1(params,Y));
emit_op2(p, OPCODE_MAX, tmp, 0, tmp, swizzle1(id,X)); /* saturate */
emit_op2(p, OPCODE_MIN, fog, WRITEMASK_X, tmp, swizzle1(id,W));
break;
}
case FOG_EXP:
emit_op2(p, OPCODE_MUL, tmp, 0, useabs ? tmp : input,
swizzle1(params,Z));
emit_op1(p, OPCODE_EX2, fog, WRITEMASK_X, ureg_negate(tmp));
break;
case FOG_EXP2:
emit_op2(p, OPCODE_MUL, tmp, 0, input, swizzle1(params,W));
emit_op2(p, OPCODE_MUL, tmp, 0, tmp, tmp);
emit_op1(p, OPCODE_EX2, fog, WRITEMASK_X, ureg_negate(tmp));
break;
}
release_temp(p, tmp);
}
else {
/* results = incoming fog coords (compute fog per-fragment later)
*
* KW: Is it really necessary to do anything in this case?
*/
emit_op1(p, useabs ? OPCODE_ABS : OPCODE_MOV, fog, 0, input);
}
}
/**
* Create a TGSI ureg_src register from a Mesa src register.
*/
static struct ureg_src
translate_src( struct st_translate *t,
const struct prog_src_register *SrcReg )
{
struct ureg_src src = src_register( t, SrcReg->File, SrcReg->Index );
src = ureg_swizzle( src,
GET_SWZ( SrcReg->Swizzle, 0 ) & 0x3,
GET_SWZ( SrcReg->Swizzle, 1 ) & 0x3,
GET_SWZ( SrcReg->Swizzle, 2 ) & 0x3,
GET_SWZ( SrcReg->Swizzle, 3 ) & 0x3);
if (SrcReg->Negate == NEGATE_XYZW)
src = ureg_negate(src);
if (SrcReg->RelAddr) {
src = ureg_src_indirect( src, ureg_src(t->address[0]));
if (SrcReg->File != PROGRAM_INPUT &&
SrcReg->File != PROGRAM_OUTPUT) {
/* If SrcReg->Index was negative, it was set to zero in
* src_register(). Reassign it now. But don't do this
* for input/output regs since they get remapped while
* const buffers don't.
*/
src.Index = SrcReg->Index;
}
}
return src;
}
/**
* Negate the value of DDY to match GL semantics where (0,0) is the
* lower-left corner of the window.
* Note that the GL_ARB_fragment_coord_conventions extension will
* effect this someday.
*/
static void emit_ddy( struct st_translate *t,
struct ureg_dst dst,
const struct prog_src_register *SrcReg )
{
struct ureg_program *ureg = t->ureg;
struct ureg_src src = translate_src( t, SrcReg );
src = ureg_negate( src );
ureg_DDY( ureg, dst, src );
}
static struct ureg calculate_light_attenuation( struct tnl_program *p,
GLuint i,
struct ureg VPpli,
struct ureg dist )
{
struct ureg attenuation = register_param3(p, STATE_LIGHT, i,
STATE_ATTENUATION);
struct ureg att = get_temp(p);
/* Calculate spot attenuation:
*/
if (!p->state->unit[i].light_spotcutoff_is_180) {
struct ureg spot_dir_norm = register_param3(p, STATE_INTERNAL,
STATE_SPOT_DIR_NORMALIZED, i);
struct ureg spot = get_temp(p);
struct ureg slt = get_temp(p);
emit_op2(p, OPCODE_DP3, spot, 0, ureg_negate(VPpli), spot_dir_norm);
emit_op2(p, OPCODE_SLT, slt, 0, swizzle1(spot_dir_norm,W), spot);
emit_op2(p, OPCODE_POW, spot, 0, spot, swizzle1(attenuation, W));
emit_op2(p, OPCODE_MUL, att, 0, slt, spot);
release_temp(p, spot);
release_temp(p, slt);
}
/* Calculate distance attenuation:
*/
if (p->state->unit[i].light_attenuated) {
/* 1/d,d,d,1/d */
emit_op1(p, OPCODE_RCP, dist, WRITEMASK_YZ, dist);
/* 1,d,d*d,1/d */
emit_op2(p, OPCODE_MUL, dist, WRITEMASK_XZ, dist, swizzle1(dist,Y));
/* 1/dist-atten */
emit_op2(p, OPCODE_DP3, dist, 0, attenuation, dist);
if (!p->state->unit[i].light_spotcutoff_is_180) {
/* dist-atten */
emit_op1(p, OPCODE_RCP, dist, 0, dist);
/* spot-atten * dist-atten */
emit_op2(p, OPCODE_MUL, att, 0, dist, att);
} else {
/* dist-atten */
emit_op1(p, OPCODE_RCP, att, 0, dist);
}
}
return att;
}
static void build_reflect_texgen( struct tnl_program *p,
struct ureg dest,
GLuint writemask )
{
struct ureg normal = get_eye_normal(p);
struct ureg eye_hat = get_eye_position_normalized(p);
struct ureg tmp = get_temp(p);
/* n.u */
emit_op2(p, OPCODE_DP3, tmp, 0, normal, eye_hat);
/* 2n.u */
emit_op2(p, OPCODE_ADD, tmp, 0, tmp, tmp);
/* (-2n.u)n + u */
emit_op3(p, OPCODE_MAD, dest, writemask, ureg_negate(tmp), normal, eye_hat);
release_temp(p, tmp);
}
static void build_sphere_texgen( struct tnl_program *p,
struct ureg dest,
GLuint writemask )
{
struct ureg normal = get_eye_normal(p);
struct ureg eye_hat = get_eye_position_normalized(p);
struct ureg tmp = get_temp(p);
struct ureg half = register_scalar_const(p, .5);
struct ureg r = get_temp(p);
struct ureg inv_m = get_temp(p);
struct ureg id = get_identity_param(p);
/* Could share the above calculations, but it would be
* a fairly odd state for someone to set (both sphere and
* reflection active for different texture coordinate
* components. Of course - if two texture units enable
* reflect and/or sphere, things start to tilt in favour
* of seperating this out:
*/
/* n.u */
emit_op2(p, OPCODE_DP3, tmp, 0, normal, eye_hat);
/* 2n.u */
emit_op2(p, OPCODE_ADD, tmp, 0, tmp, tmp);
/* (-2n.u)n + u */
emit_op3(p, OPCODE_MAD, r, 0, ureg_negate(tmp), normal, eye_hat);
/* r + 0,0,1 */
emit_op2(p, OPCODE_ADD, tmp, 0, r, swizzle(id,X,Y,W,Z));
/* rx^2 + ry^2 + (rz+1)^2 */
emit_op2(p, OPCODE_DP3, tmp, 0, tmp, tmp);
/* 2/m */
emit_op1(p, OPCODE_RSQ, tmp, 0, tmp);
/* 1/m */
emit_op2(p, OPCODE_MUL, inv_m, 0, tmp, half);
/* r/m + 1/2 */
emit_op3(p, OPCODE_MAD, dest, writemask, r, inv_m, half);
release_temp(p, tmp);
release_temp(p, r);
release_temp(p, inv_m);
}
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);
}
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_SUB(ureg, temp2, ureg_src(temp3), ureg_src(temp0));
ureg_RSQ(ureg, temp2, ureg_abs(ureg_src(temp2)));
ureg_RCP(ureg, temp2, ureg_src(temp2));
ureg_SUB(ureg, temp1, ureg_src(temp2), 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_frag_shader_weave(struct vl_compositor *c)
{
struct ureg_program *shader;
struct ureg_src i_tc[2];
struct ureg_src csc[3];
struct ureg_src sampler[3];
struct ureg_dst t_tc[2];
struct ureg_dst t_texel[2];
struct ureg_dst o_fragment;
unsigned i, j;
shader = ureg_create(TGSI_PROCESSOR_FRAGMENT);
if (!shader)
return false;
i_tc[0] = ureg_DECL_fs_input(shader, TGSI_SEMANTIC_GENERIC, VS_O_VTOP, TGSI_INTERPOLATE_LINEAR);
i_tc[1] = ureg_DECL_fs_input(shader, TGSI_SEMANTIC_GENERIC, VS_O_VBOTTOM, TGSI_INTERPOLATE_LINEAR);
for (i = 0; i < 3; ++i) {
csc[i] = ureg_DECL_constant(shader, i);
sampler[i] = ureg_DECL_sampler(shader, i);
}
for (i = 0; i < 2; ++i) {
t_tc[i] = ureg_DECL_temporary(shader);
t_texel[i] = ureg_DECL_temporary(shader);
}
o_fragment = ureg_DECL_output(shader, TGSI_SEMANTIC_COLOR, 0);
/* calculate the texture offsets
* t_tc.x = i_tc.x
* t_tc.y = (round(i_tc.y) + 0.5) / height * 2
*/
for (i = 0; i < 2; ++i) {
ureg_MOV(shader, ureg_writemask(t_tc[i], TGSI_WRITEMASK_X), i_tc[i]);
ureg_ROUND(shader, ureg_writemask(t_tc[i], TGSI_WRITEMASK_YZ), i_tc[i]);
ureg_MOV(shader, ureg_writemask(t_tc[i], TGSI_WRITEMASK_W),
ureg_imm1f(shader, i ? 0.75f : 0.25f));
ureg_ADD(shader, ureg_writemask(t_tc[i], TGSI_WRITEMASK_YZ),
ureg_src(t_tc[i]), ureg_imm1f(shader, 0.5f));
ureg_MUL(shader, ureg_writemask(t_tc[i], TGSI_WRITEMASK_Y),
ureg_src(t_tc[i]), ureg_scalar(i_tc[0], TGSI_SWIZZLE_W));
ureg_MUL(shader, ureg_writemask(t_tc[i], TGSI_WRITEMASK_Z),
ureg_src(t_tc[i]), ureg_scalar(i_tc[1], TGSI_SWIZZLE_W));
}
/* fetch the texels
* texel[0..1].x = tex(t_tc[0..1][0])
* texel[0..1].y = tex(t_tc[0..1][1])
* texel[0..1].z = tex(t_tc[0..1][2])
*/
for (i = 0; i < 2; ++i)
for (j = 0; j < 3; ++j) {
struct ureg_src src = ureg_swizzle(ureg_src(t_tc[i]),
TGSI_SWIZZLE_X, j ? TGSI_SWIZZLE_Z : TGSI_SWIZZLE_Y, TGSI_SWIZZLE_W, TGSI_SWIZZLE_W);
ureg_TEX(shader, ureg_writemask(t_texel[i], TGSI_WRITEMASK_X << j),
TGSI_TEXTURE_3D, src, sampler[j]);
}
/* calculate linear interpolation factor
* factor = |round(i_tc.y) - i_tc.y| * 2
*/
ureg_ROUND(shader, ureg_writemask(t_tc[0], TGSI_WRITEMASK_YZ), i_tc[0]);
ureg_ADD(shader, ureg_writemask(t_tc[0], TGSI_WRITEMASK_YZ),
ureg_src(t_tc[0]), ureg_negate(i_tc[0]));
ureg_MUL(shader, ureg_writemask(t_tc[0], TGSI_WRITEMASK_XY),
ureg_abs(ureg_src(t_tc[0])), ureg_imm1f(shader, 2.0f));
ureg_LRP(shader, t_texel[0], ureg_swizzle(ureg_src(t_tc[0]),
TGSI_SWIZZLE_Y, TGSI_SWIZZLE_Z, TGSI_SWIZZLE_Z, TGSI_SWIZZLE_Z),
ureg_src(t_texel[1]), ureg_src(t_texel[0]));
/* and finally do colour space transformation
* fragment = csc * texel
*/
ureg_MOV(shader, ureg_writemask(t_texel[0], TGSI_WRITEMASK_W), ureg_imm1f(shader, 1.0f));
for (i = 0; i < 3; ++i)
ureg_DP4(shader, ureg_writemask(o_fragment, TGSI_WRITEMASK_X << i), csc[i], ureg_src(t_texel[0]));
ureg_MOV(shader, ureg_writemask(o_fragment, TGSI_WRITEMASK_W), ureg_imm1f(shader, 1.0f));
for (i = 0; i < 2; ++i) {
ureg_release_temporary(shader, t_texel[i]);
ureg_release_temporary(shader, t_tc[i]);
}
ureg_END(shader);
return ureg_create_shader_and_destroy(shader, c->pipe);
}
/* Need to add some addtional parameters to allow lighting in object
* space - STATE_SPOT_DIRECTION and STATE_HALF_VECTOR implicitly assume eye
* space lighting.
*/
static void build_lighting( struct tnl_program *p )
{
const GLboolean twoside = p->state->light_twoside;
const GLboolean separate = p->state->separate_specular;
GLuint nr_lights = 0, count = 0;
struct ureg normal = get_eye_normal(p);
struct ureg lit = get_temp(p);
struct ureg dots = get_temp(p);
struct ureg _col0 = undef, _col1 = undef;
struct ureg _bfc0 = undef, _bfc1 = undef;
GLuint i;
for (i = 0; i < MAX_LIGHTS; i++)
if (p->state->unit[i].light_enabled)
nr_lights++;
set_material_flags(p);
{
struct ureg shininess = get_material(p, 0, STATE_SHININESS);
emit_op1(p, OPCODE_MOV, dots, WRITEMASK_W, swizzle1(shininess,X));
release_temp(p, shininess);
_col0 = make_temp(p, get_scenecolor(p, 0));
if (separate)
_col1 = make_temp(p, get_identity_param(p));
else
_col1 = _col0;
}
if (twoside) {
struct ureg shininess = get_material(p, 1, STATE_SHININESS);
emit_op1(p, OPCODE_MOV, dots, WRITEMASK_Z,
ureg_negate(swizzle1(shininess,X)));
release_temp(p, shininess);
_bfc0 = make_temp(p, get_scenecolor(p, 1));
if (separate)
_bfc1 = make_temp(p, get_identity_param(p));
else
_bfc1 = _bfc0;
}
/* If no lights, still need to emit the scenecolor.
*/
/* KW: changed to do this always - v1.17 "Fix lighting alpha result"?
*/
if (p->state->fragprog_inputs_read & FRAG_BIT_COL0)
{
struct ureg res0 = register_output( p, VERT_RESULT_COL0 );
emit_op1(p, OPCODE_MOV, res0, 0, _col0);
if (twoside) {
struct ureg res0 = register_output( p, VERT_RESULT_BFC0 );
emit_op1(p, OPCODE_MOV, res0, 0, _bfc0);
}
}
if (separate && (p->state->fragprog_inputs_read & FRAG_BIT_COL1)) {
struct ureg res1 = register_output( p, VERT_RESULT_COL1 );
emit_op1(p, OPCODE_MOV, res1, 0, _col1);
if (twoside) {
struct ureg res1 = register_output( p, VERT_RESULT_BFC1 );
emit_op1(p, OPCODE_MOV, res1, 0, _bfc1);
}
}
if (nr_lights == 0) {
release_temps(p);
return;
}
for (i = 0; i < MAX_LIGHTS; i++) {
if (p->state->unit[i].light_enabled) {
struct ureg half = undef;
struct ureg att = undef, VPpli = undef;
count++;
if (p->state->unit[i].light_eyepos3_is_zero) {
/* Can used precomputed constants in this case.
* Attenuation never applies to infinite lights.
*/
VPpli = register_param3(p, STATE_LIGHT, i,
STATE_POSITION_NORMALIZED);
if (p->state->light_local_viewer) {
struct ureg eye_hat = get_eye_position_normalized(p);
half = get_temp(p);
emit_op2(p, OPCODE_SUB, half, 0, VPpli, eye_hat);
emit_normalize_vec3(p, half, half);
} else {
half = register_param3(p, STATE_LIGHT, i, STATE_HALF_VECTOR);
}
}
else {
struct ureg Ppli = register_param3(p, STATE_LIGHT, i,
STATE_POSITION);
struct ureg V = get_eye_position(p);
struct ureg dist = get_temp(p);
VPpli = get_temp(p);
half = get_temp(p);
/* Calulate VPpli vector
*/
emit_op2(p, OPCODE_SUB, VPpli, 0, Ppli, V);
/* Normalize VPpli. The dist value also used in
* attenuation below.
*/
emit_op2(p, OPCODE_DP3, dist, 0, VPpli, VPpli);
emit_op1(p, OPCODE_RSQ, dist, 0, dist);
emit_op2(p, OPCODE_MUL, VPpli, 0, VPpli, dist);
/* Calculate attenuation:
*/
if (!p->state->unit[i].light_spotcutoff_is_180 ||
p->state->unit[i].light_attenuated) {
att = calculate_light_attenuation(p, i, VPpli, dist);
}
/* Calculate viewer direction, or use infinite viewer:
*/
if (p->state->light_local_viewer) {
struct ureg eye_hat = get_eye_position_normalized(p);
emit_op2(p, OPCODE_SUB, half, 0, VPpli, eye_hat);
}
else {
struct ureg z_dir = swizzle(get_identity_param(p),X,Y,W,Z);
emit_op2(p, OPCODE_ADD, half, 0, VPpli, z_dir);
}
emit_normalize_vec3(p, half, half);
release_temp(p, dist);
}
/* Calculate dot products:
*/
emit_op2(p, OPCODE_DP3, dots, WRITEMASK_X, normal, VPpli);
emit_op2(p, OPCODE_DP3, dots, WRITEMASK_Y, normal, half);
/* Front face lighting:
*/
{
struct ureg ambient = get_lightprod(p, i, 0, STATE_AMBIENT);
struct ureg diffuse = get_lightprod(p, i, 0, STATE_DIFFUSE);
struct ureg specular = get_lightprod(p, i, 0, STATE_SPECULAR);
struct ureg res0, res1;
GLuint mask0, mask1;
emit_op1(p, OPCODE_LIT, lit, 0, dots);
if (!is_undef(att))
emit_op2(p, OPCODE_MUL, lit, 0, lit, att);
mask0 = 0;
mask1 = 0;
res0 = _col0;
res1 = _col1;
if (count == nr_lights) {
if (separate) {
mask0 = WRITEMASK_XYZ;
mask1 = WRITEMASK_XYZ;
if (p->state->fragprog_inputs_read & FRAG_BIT_COL0)
res0 = register_output( p, VERT_RESULT_COL0 );
if (p->state->fragprog_inputs_read & FRAG_BIT_COL1)
res1 = register_output( p, VERT_RESULT_COL1 );
}
else {
mask1 = WRITEMASK_XYZ;
if (p->state->fragprog_inputs_read & FRAG_BIT_COL0)
res1 = register_output( p, VERT_RESULT_COL0 );
}
}
emit_op3(p, OPCODE_MAD, _col0, 0, swizzle1(lit,X), ambient, _col0);
emit_op3(p, OPCODE_MAD, res0, mask0, swizzle1(lit,Y), diffuse, _col0);
emit_op3(p, OPCODE_MAD, res1, mask1, swizzle1(lit,Z), specular, _col1);
release_temp(p, ambient);
release_temp(p, diffuse);
release_temp(p, specular);
}
/* Back face lighting:
*/
if (twoside) {
struct ureg ambient = get_lightprod(p, i, 1, STATE_AMBIENT);
struct ureg diffuse = get_lightprod(p, i, 1, STATE_DIFFUSE);
struct ureg specular = get_lightprod(p, i, 1, STATE_SPECULAR);
struct ureg res0, res1;
GLuint mask0, mask1;
emit_op1(p, OPCODE_LIT, lit, 0, ureg_negate(swizzle(dots,X,Y,W,Z)));
if (!is_undef(att))
emit_op2(p, OPCODE_MUL, lit, 0, lit, att);
mask0 = 0;
mask1 = 0;
res0 = _bfc0;
res1 = _bfc1;
if (count == nr_lights) {
if (separate) {
mask0 = WRITEMASK_XYZ;
mask1 = WRITEMASK_XYZ;
if (p->state->fragprog_inputs_read & FRAG_BIT_COL0)
res0 = register_output( p, VERT_RESULT_BFC0 );
if (p->state->fragprog_inputs_read & FRAG_BIT_COL1)
res1 = register_output( p, VERT_RESULT_BFC1 );
}
else {
mask1 = WRITEMASK_XYZ;
if (p->state->fragprog_inputs_read & FRAG_BIT_COL0)
res1 = register_output( p, VERT_RESULT_BFC0 );
}
}
emit_op3(p, OPCODE_MAD, _bfc0, 0, swizzle1(lit,X), ambient, _bfc0);
emit_op3(p, OPCODE_MAD, res0, mask0, swizzle1(lit,Y), diffuse, _bfc0);
emit_op3(p, OPCODE_MAD, res1, mask1, swizzle1(lit,Z), specular, _bfc1);
release_temp(p, ambient);
release_temp(p, diffuse);
release_temp(p, specular);
}
release_temp(p, half);
release_temp(p, VPpli);
release_temp(p, att);
}
}
release_temps( p );
}
static void *
create_ycbcr_vert_shader(struct vl_mc *r, vl_mc_ycbcr_vert_shader vs_callback, void *callback_priv)
{
struct ureg_program *shader;
struct ureg_src vrect, vpos;
struct ureg_dst t_vpos, t_vtex;
struct ureg_dst o_vpos, o_flags;
struct vertex2f scale = {
(float)VL_BLOCK_WIDTH / r->buffer_width * VL_MACROBLOCK_WIDTH / r->macroblock_size,
(float)VL_BLOCK_HEIGHT / r->buffer_height * VL_MACROBLOCK_HEIGHT / r->macroblock_size
};
unsigned label;
shader = ureg_create(TGSI_PROCESSOR_VERTEX);
if (!shader)
return NULL;
vrect = ureg_DECL_vs_input(shader, VS_I_RECT);
vpos = ureg_DECL_vs_input(shader, VS_I_VPOS);
t_vpos = calc_position(r, shader, ureg_imm2f(shader, scale.x, scale.y));
t_vtex = ureg_DECL_temporary(shader);
o_vpos = ureg_DECL_output(shader, TGSI_SEMANTIC_POSITION, VS_O_VPOS);
o_flags = ureg_DECL_output(shader, TGSI_SEMANTIC_GENERIC, VS_O_FLAGS);
/*
* o_vtex.xy = t_vpos
* o_flags.z = intra * 0.5
*
* if(interlaced) {
* t_vtex.xy = vrect.y ? { 0, scale.y } : { -scale.y : 0 }
* t_vtex.z = vpos.y % 2
* t_vtex.y = t_vtex.z ? t_vtex.x : t_vtex.y
* o_vpos.y = t_vtex.y + t_vpos.y
*
* o_flags.w = t_vtex.z ? 0 : 1
* }
*
*/
vs_callback(callback_priv, r, shader, VS_O_VTEX, t_vpos);
ureg_MUL(shader, ureg_writemask(o_flags, TGSI_WRITEMASK_Z),
ureg_scalar(vpos, TGSI_SWIZZLE_Z), ureg_imm1f(shader, 0.5f));
ureg_MOV(shader, ureg_writemask(o_flags, TGSI_WRITEMASK_W), ureg_imm1f(shader, -1.0f));
if (r->macroblock_size == VL_MACROBLOCK_HEIGHT) { //TODO
ureg_IF(shader, ureg_scalar(vpos, TGSI_SWIZZLE_W), &label);
ureg_CMP(shader, ureg_writemask(t_vtex, TGSI_WRITEMASK_XY),
ureg_negate(ureg_scalar(vrect, TGSI_SWIZZLE_Y)),
ureg_imm2f(shader, 0.0f, scale.y),
ureg_imm2f(shader, -scale.y, 0.0f));
ureg_MUL(shader, ureg_writemask(t_vtex, TGSI_WRITEMASK_Z),
ureg_scalar(vpos, TGSI_SWIZZLE_Y), ureg_imm1f(shader, 0.5f));
ureg_FRC(shader, ureg_writemask(t_vtex, TGSI_WRITEMASK_Z), ureg_src(t_vtex));
ureg_CMP(shader, ureg_writemask(t_vtex, TGSI_WRITEMASK_Y),
ureg_negate(ureg_scalar(ureg_src(t_vtex), TGSI_SWIZZLE_Z)),
ureg_scalar(ureg_src(t_vtex), TGSI_SWIZZLE_X),
ureg_scalar(ureg_src(t_vtex), TGSI_SWIZZLE_Y));
ureg_ADD(shader, ureg_writemask(o_vpos, TGSI_WRITEMASK_Y),
ureg_src(t_vpos), ureg_src(t_vtex));
ureg_CMP(shader, ureg_writemask(o_flags, TGSI_WRITEMASK_W),
ureg_negate(ureg_scalar(ureg_src(t_vtex), TGSI_SWIZZLE_Z)),
ureg_imm1f(shader, 0.0f), ureg_imm1f(shader, 1.0f));
ureg_fixup_label(shader, label, ureg_get_instruction_number(shader));
ureg_ENDIF(shader);
}
ureg_release_temporary(shader, t_vtex);
ureg_release_temporary(shader, t_vpos);
ureg_END(shader);
return ureg_create_shader_and_destroy(shader, r->pipe);
}
static void *
create_ref_frag_shader(struct vl_mc *r)
{
const float y_scale =
r->buffer_height / 2 *
r->macroblock_size / VL_MACROBLOCK_HEIGHT;
struct ureg_program *shader;
struct ureg_src tc[2], sampler;
struct ureg_dst ref, field;
struct ureg_dst fragment;
unsigned label;
shader = ureg_create(TGSI_PROCESSOR_FRAGMENT);
if (!shader)
return NULL;
tc[0] = ureg_DECL_fs_input(shader, TGSI_SEMANTIC_GENERIC, VS_O_VTOP, TGSI_INTERPOLATE_LINEAR);
tc[1] = ureg_DECL_fs_input(shader, TGSI_SEMANTIC_GENERIC, VS_O_VBOTTOM, TGSI_INTERPOLATE_LINEAR);
sampler = ureg_DECL_sampler(shader, 0);
ref = ureg_DECL_temporary(shader);
fragment = ureg_DECL_output(shader, TGSI_SEMANTIC_COLOR, 0);
field = calc_line(shader);
/*
* ref = field.z ? tc[1] : tc[0]
*
* // Adjust tc acording to top/bottom field selection
* if (|ref.z|) {
* ref.y *= y_scale
* ref.y = floor(ref.y)
* ref.y += ref.z
* ref.y /= y_scale
* }
* fragment.xyz = tex(ref, sampler[0])
*/
ureg_CMP(shader, ureg_writemask(ref, TGSI_WRITEMASK_XYZ),
ureg_negate(ureg_scalar(ureg_src(field), TGSI_SWIZZLE_Y)),
tc[1], tc[0]);
ureg_CMP(shader, ureg_writemask(fragment, TGSI_WRITEMASK_W),
ureg_negate(ureg_scalar(ureg_src(field), TGSI_SWIZZLE_Y)),
tc[1], tc[0]);
ureg_IF(shader, ureg_scalar(ureg_src(ref), TGSI_SWIZZLE_Z), &label);
ureg_MUL(shader, ureg_writemask(ref, TGSI_WRITEMASK_Y),
ureg_src(ref), ureg_imm1f(shader, y_scale));
ureg_FLR(shader, ureg_writemask(ref, TGSI_WRITEMASK_Y), ureg_src(ref));
ureg_ADD(shader, ureg_writemask(ref, TGSI_WRITEMASK_Y),
ureg_src(ref), ureg_scalar(ureg_src(ref), TGSI_SWIZZLE_Z));
ureg_MUL(shader, ureg_writemask(ref, TGSI_WRITEMASK_Y),
ureg_src(ref), ureg_imm1f(shader, 1.0f / y_scale));
ureg_fixup_label(shader, label, ureg_get_instruction_number(shader));
ureg_ENDIF(shader);
ureg_TEX(shader, ureg_writemask(fragment, TGSI_WRITEMASK_XYZ), TGSI_TEXTURE_2D, ureg_src(ref), sampler);
ureg_release_temporary(shader, ref);
ureg_release_temporary(shader, field);
ureg_END(shader);
return ureg_create_shader_and_destroy(shader, r->pipe);
}
static void *
create_mismatch_frag_shader(struct vl_idct *idct)
{
struct ureg_program *shader;
struct ureg_src addr[2];
struct ureg_dst m[8][2];
struct ureg_dst fragment;
unsigned i;
shader = ureg_create(TGSI_PROCESSOR_FRAGMENT);
if (!shader)
return NULL;
addr[0] = ureg_DECL_fs_input(shader, TGSI_SEMANTIC_GENERIC, VS_O_L_ADDR0, TGSI_INTERPOLATE_LINEAR);
addr[1] = ureg_DECL_fs_input(shader, TGSI_SEMANTIC_GENERIC, VS_O_L_ADDR1, TGSI_INTERPOLATE_LINEAR);
fragment = ureg_DECL_output(shader, TGSI_SEMANTIC_COLOR, 0);
for (i = 0; i < 8; ++i) {
m[i][0] = ureg_DECL_temporary(shader);
m[i][1] = ureg_DECL_temporary(shader);
}
for (i = 0; i < 8; ++i) {
increment_addr(shader, m[i], addr, false, false, i, idct->buffer_height);
}
for (i = 0; i < 8; ++i) {
struct ureg_src s_addr[2];
s_addr[0] = ureg_src(m[i][0]);
s_addr[1] = ureg_src(m[i][1]);
fetch_four(shader, m[i], s_addr, ureg_DECL_sampler(shader, 0), false);
}
for (i = 1; i < 8; ++i) {
ureg_ADD(shader, m[0][0], ureg_src(m[0][0]), ureg_src(m[i][0]));
ureg_ADD(shader, m[0][1], ureg_src(m[0][1]), ureg_src(m[i][1]));
}
ureg_ADD(shader, m[0][0], ureg_src(m[0][0]), ureg_src(m[0][1]));
ureg_DP4(shader, m[0][0], ureg_abs(ureg_src(m[0][0])), ureg_imm1f(shader, 1 << 14));
ureg_MUL(shader, ureg_writemask(m[0][0], TGSI_WRITEMASK_W), ureg_abs(ureg_src(m[7][1])), ureg_imm1f(shader, 1 << 14));
ureg_FRC(shader, m[0][0], ureg_src(m[0][0]));
ureg_SGT(shader, m[0][0], ureg_imm1f(shader, 0.5f), ureg_abs(ureg_src(m[0][0])));
ureg_CMP(shader, ureg_writemask(m[0][0], TGSI_WRITEMASK_W), ureg_negate(ureg_src(m[0][0])),
ureg_imm1f(shader, 1.0f / (1 << 15)), ureg_imm1f(shader, -1.0f / (1 << 15)));
ureg_MUL(shader, ureg_writemask(m[0][0], TGSI_WRITEMASK_W), ureg_src(m[0][0]),
ureg_scalar(ureg_src(m[0][0]), TGSI_SWIZZLE_X));
ureg_MOV(shader, ureg_writemask(fragment, TGSI_WRITEMASK_XYZ), ureg_src(m[7][1]));
ureg_ADD(shader, ureg_writemask(fragment, TGSI_WRITEMASK_W), ureg_src(m[0][0]), ureg_src(m[7][1]));
for (i = 0; i < 8; ++i) {
ureg_release_temporary(shader, m[i][0]);
ureg_release_temporary(shader, m[i][1]);
}
ureg_END(shader);
return ureg_create_shader_and_destroy(shader, idct->pipe);
}
static void
compile_instruction(
struct gl_context *ctx,
struct st_translate *t,
const struct prog_instruction *inst)
{
struct ureg_program *ureg = t->ureg;
GLuint i;
struct ureg_dst dst[1] = { { 0 } };
struct ureg_src src[4];
unsigned num_dst;
unsigned num_src;
num_dst = _mesa_num_inst_dst_regs( inst->Opcode );
num_src = _mesa_num_inst_src_regs( inst->Opcode );
if (num_dst)
dst[0] = translate_dst( t,
&inst->DstReg,
inst->Saturate);
for (i = 0; i < num_src; i++)
src[i] = translate_src( t, &inst->SrcReg[i] );
switch( inst->Opcode ) {
case OPCODE_SWZ:
emit_swz( t, dst[0], &inst->SrcReg[0] );
return;
case OPCODE_TEX:
case OPCODE_TXB:
case OPCODE_TXP:
src[num_src++] = t->samplers[inst->TexSrcUnit];
ureg_tex_insn( ureg,
translate_opcode( inst->Opcode ),
dst, num_dst,
st_translate_texture_target( inst->TexSrcTarget,
inst->TexShadow ),
NULL, 0,
src, num_src );
return;
case OPCODE_SCS:
dst[0] = ureg_writemask(dst[0], TGSI_WRITEMASK_XY );
ureg_insn( ureg,
translate_opcode( inst->Opcode ),
dst, num_dst,
src, num_src );
break;
case OPCODE_XPD:
dst[0] = ureg_writemask(dst[0], TGSI_WRITEMASK_XYZ );
ureg_insn( ureg,
translate_opcode( inst->Opcode ),
dst, num_dst,
src, num_src );
break;
case OPCODE_RSQ:
ureg_RSQ( ureg, dst[0], ureg_abs(src[0]) );
break;
case OPCODE_ABS:
ureg_MOV(ureg, dst[0], ureg_abs(src[0]));
break;
case OPCODE_SUB:
ureg_ADD(ureg, dst[0], src[0], ureg_negate(src[1]));
break;
default:
ureg_insn( ureg,
translate_opcode( inst->Opcode ),
dst, num_dst,
src, num_src );
break;
}
}
