i965: Add support for gl_VertexID and gl_InstanceID.

The compiler setup for these VF-uploaded attributes looks a little
cheesy with mixing system values and real VBO-sourced attributes.  It
would be nice if we could just compute the ATTR[] map to GRF index up
front and use it at visit time instead of using ir->location in the
ATTR file.  However, we don't know the reg_offset at
visit(ir_variable *) time, so we can't do the mapping that early.

Fixes piglit vertexid test.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
This commit is contained in:
Eric Anholt 2011-11-08 18:50:09 -08:00
parent 9472f66776
commit 1c65abb211
4 changed files with 63 additions and 6 deletions

View file

@ -387,6 +387,7 @@ struct brw_vs_prog_data {
const float *pull_param[MAX_UNIFORMS * 4];
bool uses_new_param_layout;
bool uses_vertexid;
};

View file

@ -591,7 +591,7 @@ static void brw_emit_vertices(struct brw_context *brw)
{
struct gl_context *ctx = &brw->intel.ctx;
struct intel_context *intel = intel_context(ctx);
GLuint i;
GLuint i, nr_elements;
brw_prepare_vertices(brw);
@ -667,17 +667,19 @@ static void brw_emit_vertices(struct brw_context *brw)
ADVANCE_BATCH();
}
nr_elements = brw->vb.nr_enabled + brw->vs.prog_data->uses_vertexid;
/* The hardware allows one more VERTEX_ELEMENTS than VERTEX_BUFFERS, presumably
* for VertexID/InstanceID.
*/
if (intel->gen >= 6) {
assert(brw->vb.nr_enabled <= 34);
assert(nr_elements <= 34);
} else {
assert(brw->vb.nr_enabled <= 18);
assert(nr_elements <= 18);
}
BEGIN_BATCH(1 + brw->vb.nr_enabled * 2);
OUT_BATCH((_3DSTATE_VERTEX_ELEMENTS << 16) | (2*brw->vb.nr_enabled - 1));
BEGIN_BATCH(1 + nr_elements * 2);
OUT_BATCH((_3DSTATE_VERTEX_ELEMENTS << 16) | (2 * nr_elements - 1));
for (i = 0; i < brw->vb.nr_enabled; i++) {
struct brw_vertex_element *input = brw->vb.enabled[i];
uint32_t format = get_surface_type(input->glarray->Type,
@ -723,6 +725,30 @@ static void brw_emit_vertices(struct brw_context *brw)
(comp3 << BRW_VE1_COMPONENT_3_SHIFT) |
((i * 4) << BRW_VE1_DST_OFFSET_SHIFT));
}
if (brw->vs.prog_data->uses_vertexid) {
uint32_t dw0 = 0, dw1 = 0;
dw1 = ((BRW_VE1_COMPONENT_STORE_VID << BRW_VE1_COMPONENT_0_SHIFT) |
(BRW_VE1_COMPONENT_STORE_IID << BRW_VE1_COMPONENT_1_SHIFT) |
(BRW_VE1_COMPONENT_STORE_PID << BRW_VE1_COMPONENT_2_SHIFT) |
(BRW_VE1_COMPONENT_STORE_0 << BRW_VE1_COMPONENT_3_SHIFT));
if (intel->gen >= 6) {
dw0 |= GEN6_VE0_VALID;
} else {
dw0 |= BRW_VE0_VALID;
dw1 |= (i * 4) << BRW_VE1_DST_OFFSET_SHIFT;
}
/* Note that for gl_VertexID, gl_InstanceID, and gl_PrimitiveID values,
* the format is ignored and the value is always int.
*/
OUT_BATCH(dw0);
OUT_BATCH(dw1);
}
CACHED_BATCH();
}

View file

@ -35,7 +35,7 @@ int
vec4_visitor::setup_attributes(int payload_reg)
{
int nr_attributes;
int attribute_map[VERT_ATTRIB_MAX];
int attribute_map[VERT_ATTRIB_MAX + 1];
nr_attributes = 0;
for (int i = 0; i < VERT_ATTRIB_MAX; i++) {
@ -45,6 +45,15 @@ vec4_visitor::setup_attributes(int payload_reg)
}
}
/* VertexID is stored by the VF as the last vertex element, but we
* don't represent it with a flag in inputs_read, so we call it
* VERT_ATTRIB_MAX.
*/
if (prog_data->uses_vertexid) {
attribute_map[VERT_ATTRIB_MAX] = payload_reg + nr_attributes;
nr_attributes++;
}
foreach_list(node, &this->instructions) {
vec4_instruction *inst = (vec4_instruction *)node;

View file

@ -880,6 +880,27 @@ vec4_visitor::visit(ir_variable *ir)
}
break;
case ir_var_system_value:
/* VertexID is stored by the VF as the last vertex element, but
* we don't represent it with a flag in inputs_read, so we call
* it VERT_ATTRIB_MAX, which setup_attributes() picks up on.
*/
reg = new(mem_ctx) dst_reg(ATTR, VERT_ATTRIB_MAX);
prog_data->uses_vertexid = true;
switch (ir->location) {
case SYSTEM_VALUE_VERTEX_ID:
reg->writemask = WRITEMASK_X;
break;
case SYSTEM_VALUE_INSTANCE_ID:
reg->writemask = WRITEMASK_Y;
break;
default:
assert(!"not reached");
break;
}
break;
default:
assert(!"not reached");
}