diff options
Diffstat (limited to 'src/mesa/shader/prog_execute.c')
| -rw-r--r-- | src/mesa/shader/prog_execute.c | 77 |
1 files changed, 45 insertions, 32 deletions
diff --git a/src/mesa/shader/prog_execute.c b/src/mesa/shader/prog_execute.c index 7f034520cd4..5641014d2c6 100644 --- a/src/mesa/shader/prog_execute.c +++ b/src/mesa/shader/prog_execute.c @@ -38,7 +38,6 @@ #include "main/glheader.h" #include "main/colormac.h" #include "main/context.h" -#include "program.h" #include "prog_execute.h" #include "prog_instruction.h" #include "prog_parameter.h" @@ -352,6 +351,28 @@ fetch_vector1(const struct prog_src_register *source, } +static GLuint +fetch_vector1ui(const struct prog_src_register *source, + const struct gl_program_machine *machine) +{ + const GLuint *src = (GLuint *) get_src_register_pointer(source, machine); + GLuint result; + + ASSERT(src); + + result = src[GET_SWZ(source->Swizzle, 0)]; + + if (source->Abs) { + result = FABSF(result); + } + if (source->Negate) { + result = -result; + } + + return result; +} + + /** * Fetch texel from texture. Use partial derivatives when possible. */ @@ -996,12 +1017,12 @@ _mesa_execute_program(GLcontext * ctx, /* XXX we could probably just use pow() here */ if (a[0] > 0.0F) { if (a[1] == 0.0 && a[3] == 0.0) - result[2] = 1.0; + result[2] = 1.0F; else result[2] = (GLfloat) _mesa_pow(a[1], a[3]); } else { - result[2] = 0.0; + result[2] = 0.0F; } result[3] = 1.0F; store_vector4(inst, machine, result); @@ -1668,13 +1689,11 @@ _mesa_execute_program(GLcontext * ctx, break; case OPCODE_UP2H: /* unpack two 16-bit floats */ { - GLfloat a[4], result[4]; - fi_type fi; - GLhalfNV hx, hy; - fetch_vector1(&inst->SrcReg[0], machine, a); - fi.f = a[0]; - hx = fi.i & 0xffff; - hy = fi.i >> 16; + const GLuint raw = fetch_vector1ui(&inst->SrcReg[0], machine); + GLfloat result[4]; + GLushort hx, hy; + hx = raw & 0xffff; + hy = raw >> 16; result[0] = result[2] = _mesa_half_to_float(hx); result[1] = result[3] = _mesa_half_to_float(hy); store_vector4(inst, machine, result); @@ -1682,13 +1701,11 @@ _mesa_execute_program(GLcontext * ctx, break; case OPCODE_UP2US: /* unpack two GLushorts */ { - GLfloat a[4], result[4]; - fi_type fi; + const GLuint raw = fetch_vector1ui(&inst->SrcReg[0], machine); + GLfloat result[4]; GLushort usx, usy; - fetch_vector1(&inst->SrcReg[0], machine, a); - fi.f = a[0]; - usx = fi.i & 0xffff; - usy = fi.i >> 16; + usx = raw & 0xffff; + usy = raw >> 16; result[0] = result[2] = usx * (1.0f / 65535.0f); result[1] = result[3] = usy * (1.0f / 65535.0f); store_vector4(inst, machine, result); @@ -1696,27 +1713,23 @@ _mesa_execute_program(GLcontext * ctx, break; case OPCODE_UP4B: /* unpack four GLbytes */ { - GLfloat a[4], result[4]; - fi_type fi; - fetch_vector1(&inst->SrcReg[0], machine, a); - fi.f = a[0]; - result[0] = (((fi.i >> 0) & 0xff) - 128) / 127.0F; - result[1] = (((fi.i >> 8) & 0xff) - 128) / 127.0F; - result[2] = (((fi.i >> 16) & 0xff) - 128) / 127.0F; - result[3] = (((fi.i >> 24) & 0xff) - 128) / 127.0F; + const GLuint raw = fetch_vector1ui(&inst->SrcReg[0], machine); + GLfloat result[4]; + result[0] = (((raw >> 0) & 0xff) - 128) / 127.0F; + result[1] = (((raw >> 8) & 0xff) - 128) / 127.0F; + result[2] = (((raw >> 16) & 0xff) - 128) / 127.0F; + result[3] = (((raw >> 24) & 0xff) - 128) / 127.0F; store_vector4(inst, machine, result); } break; case OPCODE_UP4UB: /* unpack four GLubytes */ { - GLfloat a[4], result[4]; - fi_type fi; - fetch_vector1(&inst->SrcReg[0], machine, a); - fi.f = a[0]; - result[0] = ((fi.i >> 0) & 0xff) / 255.0F; - result[1] = ((fi.i >> 8) & 0xff) / 255.0F; - result[2] = ((fi.i >> 16) & 0xff) / 255.0F; - result[3] = ((fi.i >> 24) & 0xff) / 255.0F; + const GLuint raw = fetch_vector1ui(&inst->SrcReg[0], machine); + GLfloat result[4]; + result[0] = ((raw >> 0) & 0xff) / 255.0F; + result[1] = ((raw >> 8) & 0xff) / 255.0F; + result[2] = ((raw >> 16) & 0xff) / 255.0F; + result[3] = ((raw >> 24) & 0xff) / 255.0F; store_vector4(inst, machine, result); } break; |