Merge branch 'mesa_7_7_branch'
Conflicts:
src/mesa/shader/prog_execute.c
diff --git a/src/mesa/shader/prog_execute.c b/src/mesa/shader/prog_execute.c
index 7ea792f..c212790 100644
--- a/src/mesa/shader/prog_execute.c
+++ b/src/mesa/shader/prog_execute.c
@@ -351,6 +351,28 @@
}
+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.
*/
@@ -1667,13 +1689,11 @@
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);
@@ -1681,13 +1701,11 @@
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);
@@ -1695,27 +1713,23 @@
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;