bool ETHRenderEntity::DrawProjShadow(const float maxHeight, const float minHeight, const ETHSceneProperties& sceneProps, const ETHLight& light, ETHSpriteEntity *pParent,
const bool maxOpacity, const bool drawToTarget, const float targetAngle, const Vector3& v3TargetPos)
{
if (!m_pSprite || IsHidden())
return false;
VideoPtr video = m_provider->GetVideo();
ETHShaderManagerPtr shaderManager = m_provider->GetShaderManager();
SpritePtr pShadow = shaderManager->GetProjShadow();
Vector3 v3LightPos;
Vector3 v3ParentPos(0,0,0);
const Vector3 v3EntityPos = GetPosition();
if (pParent)
{
v3ParentPos = pParent->GetPosition();
v3LightPos = Vector3(v3ParentPos.x, v3ParentPos.y, 0) + light.pos;
}
else
{
v3LightPos = light.pos;
}
// if the object is higher than the light, then the shadow shouldn't be cast on the floor
if (v3LightPos.z < v3EntityPos.z)
{
return true;
}
const float scale = (m_properties.shadowScale <= 0.0f) ? 1.0f : m_properties.shadowScale;
const float opacity = (m_properties.shadowOpacity <= 0.0f) ? 1.0f : m_properties.shadowOpacity;
const Vector2 v2Size = GetCurrentSize();
Vector2 v2ShadowSize(v2Size.x, v2Size.y);
Vector2 v2ShadowPos(v3EntityPos.x, v3EntityPos.y);
// if we are drawing to a target of a rotated entity
if (drawToTarget && targetAngle != 0)
{
// rotate the shadow position according to entity angle
Matrix4x4 matRot = RotateZ(-DegreeToRadian(targetAngle));
Vector3 newShadowPos(v2ShadowPos, 0);
newShadowPos = newShadowPos - v3TargetPos;
newShadowPos = Multiply(newShadowPos, matRot);
newShadowPos = newShadowPos + v3TargetPos;
v2ShadowPos.x = newShadowPos.x;
v2ShadowPos.y = newShadowPos.y;
// rotate the light source to cast it correctly
Vector3 newPos = v3LightPos - v3TargetPos;
newPos = Multiply(newPos, matRot);
v3LightPos = newPos + v3TargetPos;
}
Vector3 diff = v3EntityPos - v3LightPos;
const float squaredDist = DP3(diff, diff);
float squaredRange = light.range * light.range;
if (squaredDist > squaredRange)
{
return true;
}
v2ShadowSize.x *= _ETH_SHADOW_SCALEX * scale;
// calculate the correct shadow lenght according to the light height
if ((GetPosition().z+v2Size.y) < light.pos.z) // if the light is over the entity
{
const float planarDist = Distance(GetPositionXY(), ETHGlobal::ToVector2(v3LightPos));
const float verticalDist = Abs((v3EntityPos.z+v2Size.y)-v3LightPos.z);
const float totalDist = (planarDist/verticalDist)*Abs(v3LightPos.z);
v2ShadowSize.y = totalDist-planarDist;
// clamp shadow lenght to the object's height. This is not realistic
// but it looks better for the real-time shadows.
v2ShadowSize.y = Min(v2Size.y*_ETH_SHADOW_FAKE_STRETCH, v2ShadowSize.y);
}
else
{
v2ShadowSize.y *= ((drawToTarget) ? _ETH_SHADOW_SCALEY : _ETH_SHADOW_SCALEY/4);
}
// specify a minimum length for the shadow
v2ShadowSize.y = Max(v2ShadowSize.y, v2Size.y);
ShaderPtr pVS = video->GetVertexShader();
pVS->SetConstant(GS_L("shadowLength"), v2ShadowSize.y * m_properties.shadowLengthScale);
pVS->SetConstant(GS_L("entityZ"), Max(m_shadowZ, v3EntityPos.z));
pVS->SetConstant(GS_L("shadowZ"), m_shadowZ);
pVS->SetConstant(GS_L("lightPos"), v3LightPos);
video->SetSpriteDepth(
(GetType() == ETH_VERTICAL) ?
ETHGlobal::ComputeDepth(m_shadowZ, maxHeight, minHeight)
: Max(0.0f, ComputeDepth(maxHeight, minHeight) - m_layrableMinimumDepth));
v2ShadowSize.y = 1.0f;
Vector2 lightPos2(v3LightPos.x, v3LightPos.y);
const float shadowAngle = ::GetAngle((lightPos2 - Vector2(v3EntityPos.x, v3EntityPos.y))) + DegreeToRadian(targetAngle);
squaredRange = Max(squaredDist, squaredRange);
float attenBias = 1;
// adjust brightness according to ambient light
if (!maxOpacity)
{
attenBias = (1-(squaredDist/squaredRange));
//fade the color according to the light brightness
const float colorLen = Max(Max(light.color.x, light.color.y), light.color.z);
attenBias *= Min(colorLen, 1.0f);
//fade the color according to the ambient light brightness
const Vector3 &ambientColor = sceneProps.ambient;
const float ambientColorLen = 1.0f-((ambientColor.x + ambientColor.y + ambientColor.z)/3.0f);
attenBias = Min(attenBias*ambientColorLen, 1.0f);
attenBias *= Max(Min((1-(GetPosition().z/Max(GetCurrentSize().y, 1.0f))), 1.0f), 0.0f);
}
GS_BYTE alpha = static_cast<GS_BYTE>(attenBias*255.0f*opacity);
if (alpha < 8)
return true;
GS_COLOR dwShadowColor(alpha,255,255,255);
pShadow->SetOrigin(Vector2(0.5f, 0.79f));
pShadow->SetRectMode(GSRM_THREE_TRIANGLES);
pShadow->DrawShaped(v2ShadowPos, v2ShadowSize,
dwShadowColor, dwShadowColor, dwShadowColor, dwShadowColor,
RadianToDegree(shadowAngle));
return true;
}
void
gen8_vec4_generator::generate_vec4_instruction(vec4_instruction *instruction,
struct brw_reg dst,
struct brw_reg *src)
{
vec4_instruction *ir = (vec4_instruction *) instruction;
if (dst.width == BRW_WIDTH_4) {
/* This happens in attribute fixups for "dual instanced" geometry
* shaders, since they use attributes that are vec4's. Since the exec
* width is only 4, it's essential that the caller set
* force_writemask_all in order to make sure the instruction is executed
* regardless of which channels are enabled.
*/
assert(ir->force_writemask_all);
/* Fix up any <8;8,1> or <0;4,1> source registers to <4;4,1> to satisfy
* the following register region restrictions (from Graphics BSpec:
* 3D-Media-GPGPU Engine > EU Overview > Registers and Register Regions
* > Register Region Restrictions)
*
* 1. ExecSize must be greater than or equal to Width.
*
* 2. If ExecSize = Width and HorzStride != 0, VertStride must be set
* to Width * HorzStride."
*/
for (int i = 0; i < 3; i++) {
if (src[i].file == BRW_GENERAL_REGISTER_FILE)
src[i] = stride(src[i], 4, 4, 1);
}
}
switch (ir->opcode) {
case BRW_OPCODE_MOV:
MOV(dst, src[0]);
break;
case BRW_OPCODE_ADD:
ADD(dst, src[0], src[1]);
break;
case BRW_OPCODE_MUL:
MUL(dst, src[0], src[1]);
break;
case BRW_OPCODE_MACH:
MACH(dst, src[0], src[1]);
break;
case BRW_OPCODE_MAD:
MAD(dst, src[0], src[1], src[2]);
break;
case BRW_OPCODE_FRC:
FRC(dst, src[0]);
break;
case BRW_OPCODE_RNDD:
RNDD(dst, src[0]);
break;
case BRW_OPCODE_RNDE:
RNDE(dst, src[0]);
break;
case BRW_OPCODE_RNDZ:
RNDZ(dst, src[0]);
break;
case BRW_OPCODE_AND:
AND(dst, src[0], src[1]);
break;
case BRW_OPCODE_OR:
OR(dst, src[0], src[1]);
break;
case BRW_OPCODE_XOR:
XOR(dst, src[0], src[1]);
break;
case BRW_OPCODE_NOT:
NOT(dst, src[0]);
break;
case BRW_OPCODE_ASR:
ASR(dst, src[0], src[1]);
break;
case BRW_OPCODE_SHR:
SHR(dst, src[0], src[1]);
break;
case BRW_OPCODE_SHL:
SHL(dst, src[0], src[1]);
break;
case BRW_OPCODE_CMP:
CMP(dst, ir->conditional_mod, src[0], src[1]);
break;
case BRW_OPCODE_SEL:
SEL(dst, src[0], src[1]);
break;
case BRW_OPCODE_DPH:
DPH(dst, src[0], src[1]);
break;
case BRW_OPCODE_DP4:
DP4(dst, src[0], src[1]);
break;
case BRW_OPCODE_DP3:
DP3(dst, src[0], src[1]);
break;
case BRW_OPCODE_DP2:
DP2(dst, src[0], src[1]);
break;
case BRW_OPCODE_F32TO16:
/* Emulate the Gen7 zeroing bug. */
MOV(retype(dst, BRW_REGISTER_TYPE_UD), brw_imm_ud(0u));
MOV(retype(dst, BRW_REGISTER_TYPE_HF), src[0]);
break;
case BRW_OPCODE_F16TO32:
MOV(dst, retype(src[0], BRW_REGISTER_TYPE_HF));
break;
case BRW_OPCODE_LRP:
LRP(dst, src[0], src[1], src[2]);
break;
case BRW_OPCODE_BFREV:
/* BFREV only supports UD type for src and dst. */
BFREV(retype(dst, BRW_REGISTER_TYPE_UD),
retype(src[0], BRW_REGISTER_TYPE_UD));
break;
case BRW_OPCODE_FBH:
/* FBH only supports UD type for dst. */
FBH(retype(dst, BRW_REGISTER_TYPE_UD), src[0]);
break;
case BRW_OPCODE_FBL:
/* FBL only supports UD type for dst. */
FBL(retype(dst, BRW_REGISTER_TYPE_UD), src[0]);
break;
case BRW_OPCODE_CBIT:
/* CBIT only supports UD type for dst. */
CBIT(retype(dst, BRW_REGISTER_TYPE_UD), src[0]);
break;
case BRW_OPCODE_ADDC:
ADDC(dst, src[0], src[1]);
break;
case BRW_OPCODE_SUBB:
SUBB(dst, src[0], src[1]);
break;
case BRW_OPCODE_BFE:
BFE(dst, src[0], src[1], src[2]);
break;
case BRW_OPCODE_BFI1:
BFI1(dst, src[0], src[1]);
break;
case BRW_OPCODE_BFI2:
BFI2(dst, src[0], src[1], src[2]);
break;
case BRW_OPCODE_IF:
IF(ir->predicate);
break;
case BRW_OPCODE_ELSE:
ELSE();
break;
case BRW_OPCODE_ENDIF:
ENDIF();
break;
case BRW_OPCODE_DO:
DO();
break;
case BRW_OPCODE_BREAK:
BREAK();
break;
case BRW_OPCODE_CONTINUE:
CONTINUE();
break;
case BRW_OPCODE_WHILE:
WHILE();
break;
case SHADER_OPCODE_RCP:
MATH(BRW_MATH_FUNCTION_INV, dst, src[0]);
break;
case SHADER_OPCODE_RSQ:
MATH(BRW_MATH_FUNCTION_RSQ, dst, src[0]);
break;
case SHADER_OPCODE_SQRT:
MATH(BRW_MATH_FUNCTION_SQRT, dst, src[0]);
break;
case SHADER_OPCODE_EXP2:
MATH(BRW_MATH_FUNCTION_EXP, dst, src[0]);
break;
case SHADER_OPCODE_LOG2:
MATH(BRW_MATH_FUNCTION_LOG, dst, src[0]);
break;
case SHADER_OPCODE_SIN:
MATH(BRW_MATH_FUNCTION_SIN, dst, src[0]);
break;
case SHADER_OPCODE_COS:
MATH(BRW_MATH_FUNCTION_COS, dst, src[0]);
break;
case SHADER_OPCODE_POW:
MATH(BRW_MATH_FUNCTION_POW, dst, src[0], src[1]);
break;
case SHADER_OPCODE_INT_QUOTIENT:
MATH(BRW_MATH_FUNCTION_INT_DIV_QUOTIENT, dst, src[0], src[1]);
break;
case SHADER_OPCODE_INT_REMAINDER:
MATH(BRW_MATH_FUNCTION_INT_DIV_REMAINDER, dst, src[0], src[1]);
break;
case SHADER_OPCODE_TEX:
case SHADER_OPCODE_TXD:
case SHADER_OPCODE_TXF:
case SHADER_OPCODE_TXF_CMS:
case SHADER_OPCODE_TXF_MCS:
case SHADER_OPCODE_TXL:
case SHADER_OPCODE_TXS:
case SHADER_OPCODE_TG4:
case SHADER_OPCODE_TG4_OFFSET:
generate_tex(ir, dst);
break;
case VS_OPCODE_URB_WRITE:
generate_urb_write(ir, true);
break;
case SHADER_OPCODE_GEN4_SCRATCH_READ:
generate_scratch_read(ir, dst, src[0]);
break;
case SHADER_OPCODE_GEN4_SCRATCH_WRITE:
generate_scratch_write(ir, dst, src[0], src[1]);
break;
case VS_OPCODE_PULL_CONSTANT_LOAD:
case VS_OPCODE_PULL_CONSTANT_LOAD_GEN7:
generate_pull_constant_load(ir, dst, src[0], src[1]);
break;
case GS_OPCODE_URB_WRITE:
generate_urb_write(ir, false);
break;
case GS_OPCODE_THREAD_END:
generate_gs_thread_end(ir);
break;
case GS_OPCODE_SET_WRITE_OFFSET:
generate_gs_set_write_offset(dst, src[0], src[1]);
break;
case GS_OPCODE_SET_VERTEX_COUNT:
generate_gs_set_vertex_count(dst, src[0]);
break;
case GS_OPCODE_SET_DWORD_2_IMMED:
generate_gs_set_dword_2_immed(dst, src[0]);
break;
case GS_OPCODE_PREPARE_CHANNEL_MASKS:
generate_gs_prepare_channel_masks(dst);
break;
case GS_OPCODE_SET_CHANNEL_MASKS:
generate_gs_set_channel_masks(dst, src[0]);
break;
case SHADER_OPCODE_SHADER_TIME_ADD:
assert(!"XXX: Missing Gen8 vec4 support for INTEL_DEBUG=shader_time");
break;
case SHADER_OPCODE_UNTYPED_ATOMIC:
generate_untyped_atomic(ir, dst, src[0], src[1]);
break;
case SHADER_OPCODE_UNTYPED_SURFACE_READ:
generate_untyped_surface_read(ir, dst, src[0]);
break;
case VS_OPCODE_UNPACK_FLAGS_SIMD4X2:
assert(!"VS_OPCODE_UNPACK_FLAGS_SIMD4X2 should not be used on Gen8+.");
break;
default:
if (ir->opcode < (int) ARRAY_SIZE(opcode_descs)) {
_mesa_problem(ctx, "Unsupported opcode in `%s' in VS\n",
opcode_descs[ir->opcode].name);
} else {
_mesa_problem(ctx, "Unsupported opcode %d in VS", ir->opcode);
}
abort();
}
}