static void nine_ureg_tgsi_dump(struct ureg_program *ureg, boolean override)
{
if (debug_get_bool_option("NINE_FF_DUMP", FALSE) || override) {
- unsigned count;
- const struct tgsi_token *toks = ureg_get_tokens(ureg, &count);
+ const struct tgsi_token *toks = ureg_get_tokens(ureg, NULL);
tgsi_dump(toks, 0);
ureg_free_tokens(toks);
}
* position at the end.*/
ureg_MOV(ureg, tmp, vs->aVtx);
/* X from [X_min, X_min + width] to [-1, 1], same for Y. Z to [0, 1] */
- ureg_SUB(ureg, ureg_writemask(tmp, TGSI_WRITEMASK_XYZ), ureg_src(tmp), _CONST(101));
+ ureg_ADD(ureg, ureg_writemask(tmp, TGSI_WRITEMASK_XYZ), ureg_src(tmp), ureg_negate(_CONST(101)));
ureg_MUL(ureg, ureg_writemask(tmp, TGSI_WRITEMASK_XYZ), ureg_src(tmp), _CONST(100));
- ureg_SUB(ureg, ureg_writemask(tmp, TGSI_WRITEMASK_XY), ureg_src(tmp), ureg_imm1f(ureg, 1.0f));
+ ureg_ADD(ureg, ureg_writemask(tmp, TGSI_WRITEMASK_XY), ureg_src(tmp), ureg_imm1f(ureg, -1.0f));
/* Y needs to be reversed */
ureg_MOV(ureg, ureg_writemask(tmp, TGSI_WRITEMASK_Y), ureg_negate(ureg_src(tmp)));
/* inverse rhw */
for (i = 0; i < key->vertexblend; ++i) {
for (c = 0; c < 4; ++c) {
- cWM[c] = ureg_src_register(TGSI_FILE_CONSTANT, (160 + i * 4) * !key->vertexblend_indexed + c);
+ cWM[c] = ureg_src_dimension(ureg_src_register(TGSI_FILE_CONSTANT, (160 + i * 4) * !key->vertexblend_indexed + c), 0);
if (key->vertexblend_indexed)
cWM[c] = ureg_src_indirect(cWM[c], ureg_scalar(ureg_src(AR), i));
}
if (has_aNrm)
ureg_MAD(ureg, aNrm_dst, ureg_src(tmp2), ureg_scalar(vs->aWgt, i), ureg_src(aNrm_dst));
/* subtract weighted position value for last value */
- ureg_SUB(ureg, sum_blendweights, ureg_src(sum_blendweights), ureg_scalar(vs->aWgt, i));
+ ureg_ADD(ureg, sum_blendweights, ureg_src(sum_blendweights), ureg_negate(ureg_scalar(vs->aWgt, i)));
}
}
ureg_DP3(ureg, tmp_x, ureg_src(aVtx_normed), vs->aNrm);
ureg_MUL(ureg, tmp, vs->aNrm, _X(tmp));
ureg_ADD(ureg, tmp, ureg_src(tmp), ureg_src(tmp));
- ureg_SUB(ureg, ureg_writemask(input_coord, TGSI_WRITEMASK_XYZ), ureg_src(aVtx_normed), ureg_src(tmp));
+ ureg_ADD(ureg, ureg_writemask(input_coord, TGSI_WRITEMASK_XYZ), ureg_src(aVtx_normed), ureg_negate(ureg_src(tmp)));
ureg_MOV(ureg, ureg_writemask(input_coord, TGSI_WRITEMASK_W), ureg_imm1f(ureg, 1.0f));
ureg_release_temporary(ureg, aVtx_normed);
dim_input = 4;
ureg_DP3(ureg, tmp_x, ureg_src(aVtx_normed), vs->aNrm);
ureg_MUL(ureg, tmp, vs->aNrm, _X(tmp));
ureg_ADD(ureg, tmp, ureg_src(tmp), ureg_src(tmp));
- ureg_SUB(ureg, tmp, ureg_src(aVtx_normed), ureg_src(tmp));
+ ureg_ADD(ureg, tmp, ureg_src(aVtx_normed), ureg_negate(ureg_src(tmp)));
/* now tmp = normed(Vtx) - 2 dot3(normed(Vtx), Nrm) Nrm */
ureg_MOV(ureg, ureg_writemask(tmp2, TGSI_WRITEMASK_XYZ), ureg_src(tmp));
ureg_MUL(ureg, tmp2, ureg_src(tmp2), ureg_src(tmp2));
const unsigned loop_label = l++;
+ /* Declare all light constants to allow indirect adressing */
+ for (i = 32; i < 96; i++)
+ ureg_DECL_constant(ureg, i);
+
ureg_MOV(ureg, rCtr, ureg_imm1f(ureg, 32.0f)); /* &lightconst(0) */
ureg_MOV(ureg, rD, ureg_imm1f(ureg, 0.0f));
ureg_MOV(ureg, rA, ureg_imm1f(ureg, 0.0f));
/* hitDir = light.position - eyeVtx
* d = length(hitDir)
*/
- ureg_SUB(ureg, rHit, cLPos, vs->aVtx);
+ ureg_ADD(ureg, rHit, cLPos, ureg_negate(vs->aVtx));
ureg_DP3(ureg, tmp_x, ureg_src(rHit), ureg_src(rHit));
ureg_RSQ(ureg, tmp_y, _X(tmp));
ureg_MUL(ureg, tmp_x, _X(tmp), _Y(tmp)); /* length */
* spotAtt = (rho - light.cphi2) / (light.ctht2 - light.cphi2) ^ light.falloff
*/
ureg_DP3(ureg, tmp_y, ureg_negate(ureg_src(rHit)), cLDir); /* rho */
- ureg_SUB(ureg, tmp_x, _Y(tmp), cLPhi);
+ ureg_ADD(ureg, tmp_x, _Y(tmp), ureg_negate(cLPhi));
ureg_MUL(ureg, tmp_x, _X(tmp), cLSDiv);
ureg_POW(ureg, tmp_x, _X(tmp), cLFOff); /* spotAtten */
ureg_SGE(ureg, tmp_z, _Y(tmp), cLTht); /* if inside theta && phi */
if (has_aNrm) {
if (key->localviewer) {
ureg_normalize3(ureg, rMid, vs->aVtx);
- ureg_SUB(ureg, rMid, ureg_src(rHit), ureg_src(rMid));
+ ureg_ADD(ureg, rMid, ureg_src(rHit), ureg_negate(ureg_src(rMid)));
} else {
- ureg_SUB(ureg, rMid, ureg_src(rHit), ureg_imm3f(ureg, 0.0f, 0.0f, 1.0f));
+ ureg_ADD(ureg, rMid, ureg_src(rHit), ureg_imm3f(ureg, 0.0f, 0.0f, -1.0f));
}
ureg_normalize3(ureg, rMid, ureg_src(rMid));
ureg_DP3(ureg, ureg_saturate(tmp_x), vs->aNrm, ureg_src(rHit));
ureg_EX2(ureg, tmp_x, _X(tmp));
} else
if (key->fog_mode == D3DFOG_LINEAR) {
- ureg_SUB(ureg, tmp_x, _XXXX(_CONST(28)), _Z(tmp));
+ ureg_ADD(ureg, tmp_x, _XXXX(_CONST(28)), ureg_negate(_Z(tmp)));
ureg_MUL(ureg, ureg_saturate(tmp_x), _X(tmp), _YYYY(_CONST(28)));
}
ureg_MOV(ureg, oFog, _X(tmp));
ureg_END(ureg);
nine_ureg_tgsi_dump(ureg, FALSE);
- return ureg_create_shader_and_destroy(ureg, device->pipe);
+ return ureg_create_shader_and_destroy(ureg, device->context.pipe);
}
/* PS FF constants layout:
}
if (ta & D3DTA_COMPLEMENT) {
struct ureg_dst dst = ureg_DECL_temporary(ps->ureg);
- ureg_SUB(ps->ureg, dst, ureg_imm1f(ps->ureg, 1.0f), reg);
+ ureg_ADD(ps->ureg, dst, ureg_imm1f(ps->ureg, 1.0f), ureg_negate(reg));
reg = ureg_src(dst);
}
if (ta & D3DTA_ALPHAREPLICATE)
break;
case D3DTOP_ADDSIGNED:
ureg_ADD(ureg, tmp, arg[1], arg[2]);
- ureg_SUB(ureg, dst, ureg_src(tmp), ureg_imm1f(ureg, 0.5f));
+ ureg_ADD(ureg, dst, ureg_src(tmp), ureg_imm1f(ureg, -0.5f));
break;
case D3DTOP_ADDSIGNED2X:
ureg_ADD(ureg, tmp, arg[1], arg[2]);
ureg_MAD(ureg, dst, ureg_src(tmp), ureg_imm1f(ureg, 2.0f), ureg_imm1f(ureg, -1.0f));
break;
case D3DTOP_SUBTRACT:
- ureg_SUB(ureg, dst, arg[1], arg[2]);
+ ureg_ADD(ureg, dst, arg[1], ureg_negate(arg[2]));
break;
case D3DTOP_ADDSMOOTH:
- ureg_SUB(ureg, tmp, ureg_imm1f(ureg, 1.0f), arg[1]);
+ ureg_ADD(ureg, tmp, ureg_imm1f(ureg, 1.0f), ureg_negate(arg[1]));
ureg_MAD(ureg, dst, ureg_src(tmp), arg[2], arg[1]);
break;
case D3DTOP_BLENDDIFFUSEALPHA:
ureg_LRP(ureg, dst, _WWWW(_CONST(20)), arg[1], arg[2]);
break;
case D3DTOP_BLENDTEXTUREALPHAPM:
- ureg_SUB(ureg, tmp_x, ureg_imm1f(ureg, 1.0f), _W(ps->rTex));
+ ureg_ADD(ureg, tmp_x, ureg_imm1f(ureg, 1.0f), ureg_negate(_W(ps->rTex)));
ureg_MAD(ureg, dst, arg[2], _X(tmp), arg[1]);
break;
case D3DTOP_BLENDCURRENTALPHA:
ureg_MAD(ureg, dst, arg[1], arg[2], _WWWW(arg[1]));
break;
case D3DTOP_MODULATEINVALPHA_ADDCOLOR:
- ureg_SUB(ureg, tmp_x, ureg_imm1f(ureg, 1.0f), _WWWW(arg[1]));
+ ureg_ADD(ureg, tmp_x, ureg_imm1f(ureg, 1.0f), ureg_negate(_WWWW(arg[1])));
ureg_MAD(ureg, dst, _X(tmp), arg[2], arg[1]);
break;
case D3DTOP_MODULATEINVCOLOR_ADDALPHA:
- ureg_SUB(ureg, tmp, ureg_imm1f(ureg, 1.0f), arg[1]);
+ ureg_ADD(ureg, tmp, ureg_imm1f(ureg, 1.0f), ureg_negate(arg[1]));
ureg_MAD(ureg, dst, ureg_src(tmp), arg[2], _WWWW(arg[1]));
break;
case D3DTOP_BUMPENVMAP:
case D3DTOP_BUMPENVMAPLUMINANCE:
break;
case D3DTOP_DOTPRODUCT3:
- ureg_SUB(ureg, tmp, arg[1], ureg_imm4f(ureg,0.5,0.5,0.5,0.5));
- ureg_SUB(ureg, tmp2, arg[2] , ureg_imm4f(ureg,0.5,0.5,0.5,0.5));
+ ureg_ADD(ureg, tmp, arg[1], ureg_imm4f(ureg,-0.5,-0.5,-0.5,-0.5));
+ ureg_ADD(ureg, tmp2, arg[2] , ureg_imm4f(ureg,-0.5,-0.5,-0.5,-0.5));
ureg_DP3(ureg, tmp, ureg_src(tmp), ureg_src(tmp2));
ureg_MUL(ureg, ureg_saturate(dst), ureg_src(tmp), ureg_imm4f(ureg,4.0,4.0,4.0,4.0));
break;
ureg_MUL(ureg, ps.rMod, ps.rCurSrc, ps.rTexSrc);
}
- colorarg[0] = (key->ts[s].colorarg0 | ((key->colorarg_b4[0] >> s) << 4) | ((key->colorarg_b5[0] >> s) << 5)) & 0x3f;
- colorarg[1] = (key->ts[s].colorarg1 | ((key->colorarg_b4[1] >> s) << 4) | ((key->colorarg_b5[1] >> s) << 5)) & 0x3f;
- colorarg[2] = (key->ts[s].colorarg2 | ((key->colorarg_b4[2] >> s) << 4) | ((key->colorarg_b5[2] >> s) << 5)) & 0x3f;
+ colorarg[0] = (key->ts[s].colorarg0 | (((key->colorarg_b4[0] >> s) & 0x1) << 4) | ((key->colorarg_b5[0] >> s) << 5)) & 0x3f;
+ colorarg[1] = (key->ts[s].colorarg1 | (((key->colorarg_b4[1] >> s) & 0x1) << 4) | ((key->colorarg_b5[1] >> s) << 5)) & 0x3f;
+ colorarg[2] = (key->ts[s].colorarg2 | (((key->colorarg_b4[2] >> s) & 0x1) << 4) | ((key->colorarg_b5[2] >> s) << 5)) & 0x3f;
alphaarg[0] = (key->ts[s].alphaarg0 | ((key->alphaarg_b4[0] >> s) << 4)) & 0x1f;
alphaarg[1] = (key->ts[s].alphaarg1 | ((key->alphaarg_b4[1] >> s) << 4)) & 0x1f;
alphaarg[2] = (key->ts[s].alphaarg2 | ((key->alphaarg_b4[2] >> s) << 4)) & 0x1f;
ureg_EX2(ureg, rFog, _X(rFog));
} else
if (key->fog_mode == D3DFOG_LINEAR) {
- ureg_SUB(ureg, rFog, _XXXX(_CONST(22)), _X(rFog));
+ ureg_ADD(ureg, rFog, _XXXX(_CONST(22)), ureg_negate(_X(rFog)));
ureg_MUL(ureg, ureg_saturate(rFog), _X(rFog), _YYYY(_CONST(22)));
}
ureg_LRP(ureg, ureg_writemask(oCol, TGSI_WRITEMASK_XYZ), _X(rFog), ps.rCurSrc, _CONST(21));
ureg_END(ureg);
nine_ureg_tgsi_dump(ureg, FALSE);
- return ureg_create_shader_and_destroy(ureg, device->pipe);
+ return ureg_create_shader_and_destroy(ureg, device->context.pipe);
}
static struct NineVertexShader9 *
nine_ff_get_vs(struct NineDevice9 *device)
{
- const struct nine_state *state = &device->state;
const struct nine_context *context = &device->context;
struct NineVertexShader9 *vs;
enum pipe_error err;
key.passthrough = 0;
key.pointscale = !!context->rs[D3DRS_POINTSCALEENABLE];
- key.lighting = !!context->rs[D3DRS_LIGHTING] && state->ff.num_lights_active;
- key.darkness = !!context->rs[D3DRS_LIGHTING] && !state->ff.num_lights_active;
+ key.lighting = !!context->rs[D3DRS_LIGHTING] && context->ff.num_lights_active;
+ key.darkness = !!context->rs[D3DRS_LIGHTING] && !context->ff.num_lights_active;
if (key.position_t) {
key.darkness = 0; /* |= key.lighting; */ /* XXX ? */
key.lighting = 0;
}
for (s = 0; s < 8; ++s) {
- unsigned gen = (state->ff.tex_stage[s][D3DTSS_TEXCOORDINDEX] >> 16) + 1;
- unsigned idx = state->ff.tex_stage[s][D3DTSS_TEXCOORDINDEX] & 7;
+ unsigned gen = (context->ff.tex_stage[s][D3DTSS_TEXCOORDINDEX] >> 16) + 1;
+ unsigned idx = context->ff.tex_stage[s][D3DTSS_TEXCOORDINDEX] & 7;
unsigned dim;
if (key.position_t && gen > NINED3DTSS_TCI_PASSTHRU)
key.tc_idx |= idx << (s * 3);
key.tc_dim_input |= ((input_texture_coord[idx]-1) & 0x3) << (s * 2);
- dim = state->ff.tex_stage[s][D3DTSS_TEXTURETRANSFORMFLAGS] & 0x7;
+ dim = context->ff.tex_stage[s][D3DTSS_TEXTURETRANSFORMFLAGS] & 0x7;
if (dim > 4)
dim = input_texture_coord[idx];
if (dim == 1) /* NV behaviour */
key.tc_dim_output |= dim << (s * 3);
}
+ DBG("VS ff key hash: %x\n", nine_ff_vs_key_hash(&key));
vs = util_hash_table_get(device->ff.ht_vs, &key);
if (vs)
return vs;
(void)err;
assert(err == PIPE_OK);
device->ff.num_vs++;
- NineUnknown_ConvertRefToBind(NineUnknown(vs));
vs->num_inputs = bld.num_inputs;
for (n = 0; n < bld.num_inputs; ++n)
return vs;
}
-#define GET_D3DTS(n) nine_state_access_transform(&state->ff, D3DTS_##n, FALSE)
+#define GET_D3DTS(n) nine_state_access_transform(&context->ff, D3DTS_##n, FALSE)
#define IS_D3DTS_DIRTY(s,n) ((s)->ff.changed.transform[(D3DTS_##n) / 32] & (1 << ((D3DTS_##n) % 32)))
static struct NinePixelShader9 *
nine_ff_get_ps(struct NineDevice9 *device)
{
- struct nine_state *state = &device->state;
struct nine_context *context = &device->context;
D3DMATRIX *projection_matrix = GET_D3DTS(PROJECTION);
struct NinePixelShader9 *ps;
memset(&key, 0, sizeof(key));
for (s = 0; s < 8; ++s) {
- key.ts[s].colorop = state->ff.tex_stage[s][D3DTSS_COLOROP];
- key.ts[s].alphaop = state->ff.tex_stage[s][D3DTSS_ALPHAOP];
+ key.ts[s].colorop = context->ff.tex_stage[s][D3DTSS_COLOROP];
+ key.ts[s].alphaop = context->ff.tex_stage[s][D3DTSS_ALPHAOP];
const uint8_t used_c = ps_d3dtop_args_mask(key.ts[s].colorop);
const uint8_t used_a = ps_d3dtop_args_mask(key.ts[s].alphaop);
/* MSDN says D3DTOP_DISABLE disables this and all subsequent stages.
break;
}
- if (!context->texture[s] &&
- ((state->ff.tex_stage[s][D3DTSS_COLORARG0] == D3DTA_TEXTURE &&
+ if (!context->texture[s].enabled &&
+ ((context->ff.tex_stage[s][D3DTSS_COLORARG0] == D3DTA_TEXTURE &&
used_c & 0x1) ||
- (state->ff.tex_stage[s][D3DTSS_COLORARG1] == D3DTA_TEXTURE &&
+ (context->ff.tex_stage[s][D3DTSS_COLORARG1] == D3DTA_TEXTURE &&
used_c & 0x2) ||
- (state->ff.tex_stage[s][D3DTSS_COLORARG2] == D3DTA_TEXTURE &&
+ (context->ff.tex_stage[s][D3DTSS_COLORARG2] == D3DTA_TEXTURE &&
used_c & 0x4))) {
/* Tested on Windows: Invalid texture read disables the stage
* and the subsequent ones, but only for colorop. For alpha,
break;
}
- if (state->ff.tex_stage[s][D3DTSS_COLORARG0] == D3DTA_TEXTURE ||
- state->ff.tex_stage[s][D3DTSS_COLORARG1] == D3DTA_TEXTURE ||
- state->ff.tex_stage[s][D3DTSS_COLORARG2] == D3DTA_TEXTURE ||
- state->ff.tex_stage[s][D3DTSS_ALPHAARG0] == D3DTA_TEXTURE ||
- state->ff.tex_stage[s][D3DTSS_ALPHAARG1] == D3DTA_TEXTURE ||
- state->ff.tex_stage[s][D3DTSS_ALPHAARG2] == D3DTA_TEXTURE)
+ if (context->ff.tex_stage[s][D3DTSS_COLORARG0] == D3DTA_TEXTURE ||
+ context->ff.tex_stage[s][D3DTSS_COLORARG1] == D3DTA_TEXTURE ||
+ context->ff.tex_stage[s][D3DTSS_COLORARG2] == D3DTA_TEXTURE ||
+ context->ff.tex_stage[s][D3DTSS_ALPHAARG0] == D3DTA_TEXTURE ||
+ context->ff.tex_stage[s][D3DTSS_ALPHAARG1] == D3DTA_TEXTURE ||
+ context->ff.tex_stage[s][D3DTSS_ALPHAARG2] == D3DTA_TEXTURE)
sampler_mask |= (1 << s);
if (key.ts[s].colorop != D3DTOP_DISABLE) {
- if (used_c & 0x1) key.ts[s].colorarg0 = state->ff.tex_stage[s][D3DTSS_COLORARG0];
- if (used_c & 0x2) key.ts[s].colorarg1 = state->ff.tex_stage[s][D3DTSS_COLORARG1];
- if (used_c & 0x4) key.ts[s].colorarg2 = state->ff.tex_stage[s][D3DTSS_COLORARG2];
- if (used_c & 0x1) key.colorarg_b4[0] |= (state->ff.tex_stage[s][D3DTSS_COLORARG0] >> 4) << s;
- if (used_c & 0x1) key.colorarg_b5[0] |= (state->ff.tex_stage[s][D3DTSS_COLORARG0] >> 5) << s;
- if (used_c & 0x2) key.colorarg_b4[1] |= (state->ff.tex_stage[s][D3DTSS_COLORARG1] >> 4) << s;
- if (used_c & 0x2) key.colorarg_b5[1] |= (state->ff.tex_stage[s][D3DTSS_COLORARG1] >> 5) << s;
- if (used_c & 0x4) key.colorarg_b4[2] |= (state->ff.tex_stage[s][D3DTSS_COLORARG2] >> 4) << s;
- if (used_c & 0x4) key.colorarg_b5[2] |= (state->ff.tex_stage[s][D3DTSS_COLORARG2] >> 5) << s;
+ if (used_c & 0x1) key.ts[s].colorarg0 = context->ff.tex_stage[s][D3DTSS_COLORARG0] & 0x7;
+ if (used_c & 0x2) key.ts[s].colorarg1 = context->ff.tex_stage[s][D3DTSS_COLORARG1] & 0x7;
+ if (used_c & 0x4) key.ts[s].colorarg2 = context->ff.tex_stage[s][D3DTSS_COLORARG2] & 0x7;
+ if (used_c & 0x1) key.colorarg_b4[0] |= ((context->ff.tex_stage[s][D3DTSS_COLORARG0] >> 4) & 0x1) << s;
+ if (used_c & 0x1) key.colorarg_b5[0] |= ((context->ff.tex_stage[s][D3DTSS_COLORARG0] >> 5) & 0x1) << s;
+ if (used_c & 0x2) key.colorarg_b4[1] |= ((context->ff.tex_stage[s][D3DTSS_COLORARG1] >> 4) & 0x1) << s;
+ if (used_c & 0x2) key.colorarg_b5[1] |= ((context->ff.tex_stage[s][D3DTSS_COLORARG1] >> 5) & 0x1) << s;
+ if (used_c & 0x4) key.colorarg_b4[2] |= ((context->ff.tex_stage[s][D3DTSS_COLORARG2] >> 4) & 0x1) << s;
+ if (used_c & 0x4) key.colorarg_b5[2] |= ((context->ff.tex_stage[s][D3DTSS_COLORARG2] >> 5) & 0x1) << s;
}
if (key.ts[s].alphaop != D3DTOP_DISABLE) {
- if (used_a & 0x1) key.ts[s].alphaarg0 = state->ff.tex_stage[s][D3DTSS_ALPHAARG0];
- if (used_a & 0x2) key.ts[s].alphaarg1 = state->ff.tex_stage[s][D3DTSS_ALPHAARG1];
- if (used_a & 0x4) key.ts[s].alphaarg2 = state->ff.tex_stage[s][D3DTSS_ALPHAARG2];
- if (used_a & 0x1) key.alphaarg_b4[0] |= (state->ff.tex_stage[s][D3DTSS_ALPHAARG0] >> 4) << s;
- if (used_a & 0x2) key.alphaarg_b4[1] |= (state->ff.tex_stage[s][D3DTSS_ALPHAARG1] >> 4) << s;
- if (used_a & 0x4) key.alphaarg_b4[2] |= (state->ff.tex_stage[s][D3DTSS_ALPHAARG2] >> 4) << s;
+ if (used_a & 0x1) key.ts[s].alphaarg0 = context->ff.tex_stage[s][D3DTSS_ALPHAARG0] & 0x7;
+ if (used_a & 0x2) key.ts[s].alphaarg1 = context->ff.tex_stage[s][D3DTSS_ALPHAARG1] & 0x7;
+ if (used_a & 0x4) key.ts[s].alphaarg2 = context->ff.tex_stage[s][D3DTSS_ALPHAARG2] & 0x7;
+ if (used_a & 0x1) key.alphaarg_b4[0] |= ((context->ff.tex_stage[s][D3DTSS_ALPHAARG0] >> 4) & 0x1) << s;
+ if (used_a & 0x2) key.alphaarg_b4[1] |= ((context->ff.tex_stage[s][D3DTSS_ALPHAARG1] >> 4) & 0x1) << s;
+ if (used_a & 0x4) key.alphaarg_b4[2] |= ((context->ff.tex_stage[s][D3DTSS_ALPHAARG2] >> 4) & 0x1) << s;
}
- key.ts[s].resultarg = state->ff.tex_stage[s][D3DTSS_RESULTARG] == D3DTA_TEMP;
+ key.ts[s].resultarg = context->ff.tex_stage[s][D3DTSS_RESULTARG] == D3DTA_TEMP;
- if (context->texture[s]) {
- switch (context->texture[s]->base.type) {
+ if (context->texture[s].enabled) {
+ switch (context->texture[s].type) {
case D3DRTYPE_TEXTURE: key.ts[s].textarget = 1; break;
case D3DRTYPE_VOLUMETEXTURE: key.ts[s].textarget = 2; break;
case D3DRTYPE_CUBETEXTURE: key.ts[s].textarget = 3; break;
if (s >= 1)
key.ts[s-1].resultarg = 0;
- key.projected = nine_ff_get_projected_key(state, context);
+ key.projected = nine_ff_get_projected_key(context);
key.specular = !!context->rs[D3DRS_SPECULARENABLE];
for (; s < 8; ++s)
!(projection_matrix->_34 == 0.0f &&
projection_matrix->_44 == 1.0f);
+ DBG("PS ff key hash: %x\n", nine_ff_ps_key_hash(&key));
ps = util_hash_table_get(device->ff.ht_ps, &key);
if (ps)
return ps;
(void)err;
assert(err == PIPE_OK);
device->ff.num_ps++;
- NineUnknown_ConvertRefToBind(NineUnknown(ps));
ps->rt_mask = 0x1;
ps->sampler_mask = sampler_mask;
static void
nine_ff_load_vs_transforms(struct NineDevice9 *device)
{
- struct nine_state *state = &device->state;
struct nine_context *context = &device->context;
D3DMATRIX T;
D3DMATRIX *M = (D3DMATRIX *)device->ff.vs_const;
/* TODO: make this nicer, and only upload the ones we need */
/* TODO: use ff.vs_const as storage of W, V, P matrices */
- if (IS_D3DTS_DIRTY(state, WORLD) ||
- IS_D3DTS_DIRTY(state, VIEW) ||
- IS_D3DTS_DIRTY(state, PROJECTION)) {
+ if (IS_D3DTS_DIRTY(context, WORLD) ||
+ IS_D3DTS_DIRTY(context, VIEW) ||
+ IS_D3DTS_DIRTY(context, PROJECTION)) {
/* WVP, WV matrices */
nine_d3d_matrix_matrix_mul(&M[1], GET_D3DTS(WORLD), GET_D3DTS(VIEW));
nine_d3d_matrix_matrix_mul(&M[0], &M[1], GET_D3DTS(PROJECTION));
static void
nine_ff_load_lights(struct NineDevice9 *device)
{
- struct nine_state *state = &device->state;
struct nine_context *context = &device->context;
struct fvec4 *dst = (struct fvec4 *)device->ff.vs_const;
unsigned l;
- if (state->changed.group & NINE_STATE_FF_MATERIAL) {
- const D3DMATERIAL9 *mtl = &state->ff.material;
+ if (context->changed.group & NINE_STATE_FF_MATERIAL) {
+ const D3DMATERIAL9 *mtl = &context->ff.material;
memcpy(&dst[20], &mtl->Diffuse, 4 * sizeof(float));
memcpy(&dst[21], &mtl->Ambient, 4 * sizeof(float));
dst[19].z = dst[25].z * mtl->Ambient.b + mtl->Emissive.b;
}
- if (!(state->changed.group & NINE_STATE_FF_LIGHTING))
+ if (!(context->changed.group & NINE_STATE_FF_LIGHTING))
return;
- for (l = 0; l < state->ff.num_lights_active; ++l) {
- const D3DLIGHT9 *light = &state->ff.light[state->ff.active_light[l]];
+ for (l = 0; l < context->ff.num_lights_active; ++l) {
+ const D3DLIGHT9 *light = &context->ff.light[context->ff.active_light[l]];
dst[32 + l * 8].x = light->Type;
dst[32 + l * 8].y = light->Attenuation0;
dst[38 + l * 8].x = cosf(light->Theta * 0.5f);
dst[38 + l * 8].y = cosf(light->Phi * 0.5f);
dst[38 + l * 8].z = 1.0f / (dst[38 + l * 8].x - dst[38 + l * 8].y);
- dst[39 + l * 8].w = (l + 1) == state->ff.num_lights_active;
+ dst[39 + l * 8].w = (float)((l + 1) == context->ff.num_lights_active);
}
}
static void
nine_ff_load_point_and_fog_params(struct NineDevice9 *device)
{
- const struct nine_state *state = &device->state;
struct nine_context *context = &device->context;
struct fvec4 *dst = (struct fvec4 *)device->ff.vs_const;
- if (!(state->changed.group & NINE_STATE_FF_OTHER))
+ if (!(context->changed.group & NINE_STATE_FF_VS_OTHER))
return;
dst[26].x = asfloat(context->rs[D3DRS_POINTSIZE_MIN]);
dst[26].y = asfloat(context->rs[D3DRS_POINTSIZE_MAX]);
static void
nine_ff_load_tex_matrices(struct NineDevice9 *device)
{
- struct nine_state *state = &device->state;
+ struct nine_context *context = &device->context;
D3DMATRIX *M = (D3DMATRIX *)device->ff.vs_const;
unsigned s;
- if (!(state->ff.changed.transform[0] & 0xff0000))
+ if (!(context->ff.changed.transform[0] & 0xff0000))
return;
for (s = 0; s < 8; ++s) {
- if (IS_D3DTS_DIRTY(state, TEXTURE0 + s))
- nine_d3d_matrix_transpose(&M[32 + s], nine_state_access_transform(&state->ff, D3DTS_TEXTURE0 + s, FALSE));
+ if (IS_D3DTS_DIRTY(context, TEXTURE0 + s))
+ nine_d3d_matrix_transpose(&M[32 + s], nine_state_access_transform(&context->ff, D3DTS_TEXTURE0 + s, FALSE));
}
}
static void
nine_ff_load_ps_params(struct NineDevice9 *device)
{
- const struct nine_state *state = &device->state;
struct nine_context *context = &device->context;
struct fvec4 *dst = (struct fvec4 *)device->ff.ps_const;
unsigned s;
- if (!(state->changed.group & (NINE_STATE_FF_PSSTAGES | NINE_STATE_FF_OTHER)))
+ if (!(context->changed.group & NINE_STATE_FF_PS_CONSTS))
return;
for (s = 0; s < 8; ++s)
- d3dcolor_to_rgba(&dst[s].x, state->ff.tex_stage[s][D3DTSS_CONSTANT]);
+ d3dcolor_to_rgba(&dst[s].x, context->ff.tex_stage[s][D3DTSS_CONSTANT]);
for (s = 0; s < 8; ++s) {
- dst[8 + s].x = asfloat(state->ff.tex_stage[s][D3DTSS_BUMPENVMAT00]);
- dst[8 + s].y = asfloat(state->ff.tex_stage[s][D3DTSS_BUMPENVMAT01]);
- dst[8 + s].z = asfloat(state->ff.tex_stage[s][D3DTSS_BUMPENVMAT10]);
- dst[8 + s].w = asfloat(state->ff.tex_stage[s][D3DTSS_BUMPENVMAT11]);
+ dst[8 + s].x = asfloat(context->ff.tex_stage[s][D3DTSS_BUMPENVMAT00]);
+ dst[8 + s].y = asfloat(context->ff.tex_stage[s][D3DTSS_BUMPENVMAT01]);
+ dst[8 + s].z = asfloat(context->ff.tex_stage[s][D3DTSS_BUMPENVMAT10]);
+ dst[8 + s].w = asfloat(context->ff.tex_stage[s][D3DTSS_BUMPENVMAT11]);
if (s & 1) {
- dst[16 + s / 2].z = asfloat(state->ff.tex_stage[s][D3DTSS_BUMPENVLSCALE]);
- dst[16 + s / 2].w = asfloat(state->ff.tex_stage[s][D3DTSS_BUMPENVLOFFSET]);
+ dst[16 + s / 2].z = asfloat(context->ff.tex_stage[s][D3DTSS_BUMPENVLSCALE]);
+ dst[16 + s / 2].w = asfloat(context->ff.tex_stage[s][D3DTSS_BUMPENVLOFFSET]);
} else {
- dst[16 + s / 2].x = asfloat(state->ff.tex_stage[s][D3DTSS_BUMPENVLSCALE]);
- dst[16 + s / 2].y = asfloat(state->ff.tex_stage[s][D3DTSS_BUMPENVLOFFSET]);
+ dst[16 + s / 2].x = asfloat(context->ff.tex_stage[s][D3DTSS_BUMPENVLSCALE]);
+ dst[16 + s / 2].y = asfloat(context->ff.tex_stage[s][D3DTSS_BUMPENVLOFFSET]);
}
}
void
nine_ff_update(struct NineDevice9 *device)
{
- struct nine_state *state = &device->state;
struct nine_context *context = &device->context;
struct pipe_constant_buffer cb;
- DBG("vs=%p ps=%p\n", context->vs, device->state.ps);
+ DBG("vs=%p ps=%p\n", context->vs, context->ps);
/* NOTE: the only reference belongs to the hash table */
if (!context->programmable_vs) {
device->ff.vs = nine_ff_get_vs(device);
- device->state.changed.group |= NINE_STATE_VS;
+ context->changed.group |= NINE_STATE_VS;
}
- if (!device->state.ps) {
+ if (!context->ps) {
device->ff.ps = nine_ff_get_ps(device);
- device->state.changed.group |= NINE_STATE_PS;
+ context->changed.group |= NINE_STATE_PS;
}
if (!context->programmable_vs) {
nine_ff_load_point_and_fog_params(device);
nine_ff_load_viewport_info(device);
- memset(state->ff.changed.transform, 0, sizeof(state->ff.changed.transform));
+ memset(context->ff.changed.transform, 0, sizeof(context->ff.changed.transform));
cb.buffer_offset = 0;
cb.buffer = NULL;
cb.user_buffer = device->ff.vs_const;
cb.buffer_size = NINE_FF_NUM_VS_CONST * 4 * sizeof(float);
- if (!device->driver_caps.user_cbufs) {
- context->pipe.cb_vs_ff.buffer_size = cb.buffer_size;
- u_upload_data(device->constbuf_uploader,
- 0,
- cb.buffer_size,
- device->constbuf_alignment,
- cb.user_buffer,
- &context->pipe.cb_vs_ff.buffer_offset,
- &context->pipe.cb_vs_ff.buffer);
- u_upload_unmap(device->constbuf_uploader);
- context->pipe.cb_vs_ff.user_buffer = NULL;
- } else
- context->pipe.cb_vs_ff = cb;
+ context->pipe_data.cb_vs_ff = cb;
context->commit |= NINE_STATE_COMMIT_CONST_VS;
+
+ context->changed.group &= ~NINE_STATE_FF_VS;
}
- if (!device->state.ps) {
+ if (!context->ps) {
nine_ff_load_ps_params(device);
cb.buffer_offset = 0;
cb.user_buffer = device->ff.ps_const;
cb.buffer_size = NINE_FF_NUM_PS_CONST * 4 * sizeof(float);
- if (!device->driver_caps.user_cbufs) {
- context->pipe.cb_ps_ff.buffer_size = cb.buffer_size;
- u_upload_data(device->constbuf_uploader,
- 0,
- cb.buffer_size,
- device->constbuf_alignment,
- cb.user_buffer,
- &context->pipe.cb_ps_ff.buffer_offset,
- &context->pipe.cb_ps_ff.buffer);
- u_upload_unmap(device->constbuf_uploader);
- context->pipe.cb_ps_ff.user_buffer = NULL;
- } else
- context->pipe.cb_ps_ff = cb;
+ context->pipe_data.cb_ps_ff = cb;
context->commit |= NINE_STATE_COMMIT_CONST_PS;
- }
- device->state.changed.group &= ~NINE_STATE_FF;
+ context->changed.group &= ~NINE_STATE_FF_PS;
+ }
}
static void
nine_ff_prune_vs(struct NineDevice9 *device)
{
+ struct nine_context *context = &device->context;
+
if (device->ff.num_vs > 100) {
/* could destroy the bound one here, so unbind */
- device->pipe->bind_vs_state(device->pipe, NULL);
+ context->pipe->bind_vs_state(context->pipe, NULL);
util_hash_table_foreach(device->ff.ht_vs, nine_ff_ht_delete_cb, NULL);
util_hash_table_clear(device->ff.ht_vs);
device->ff.num_vs = 0;
- device->state.changed.group |= NINE_STATE_VS;
+ context->changed.group |= NINE_STATE_VS;
}
}
static void
nine_ff_prune_ps(struct NineDevice9 *device)
{
+ struct nine_context *context = &device->context;
+
if (device->ff.num_ps > 100) {
/* could destroy the bound one here, so unbind */
- device->pipe->bind_fs_state(device->pipe, NULL);
+ context->pipe->bind_fs_state(context->pipe, NULL);
util_hash_table_foreach(device->ff.ht_ps, nine_ff_ht_delete_cb, NULL);
util_hash_table_clear(device->ff.ht_ps);
device->ff.num_ps = 0;
- device->state.changed.group |= NINE_STATE_PS;
+ context->changed.group |= NINE_STATE_PS;
}
}
M->m[2][0] * D->m[0][2] +
M->m[3][0] * D->m[0][3];
- if (det < 1e-30) {/* non inversible */
+ if (fabsf(det) < 1e-30) {/* non inversible */
*D = *M; /* wine tests */
return;
}
projects / mesa.git / blobdiff