i965: Don't re-layout varyings for separate shader programs.

Previously, our VUE map code always assigned slots to varyings
sequentially, in one contiguous block.

This was a bad fit for separate shaders - the GS input layout depended
or the VS output layout, so if we swapped out vertex shaders, we might
have to recompile the GS on the fly - which rather defeats the point of
using separate shader objects.  (Tessellation would suffer from this
as well - we could have to recompile the HS, DS, and GS.)

Instead, this patch makes the VUE map for separate shaders use a fixed
layout, based on the input/output variable's location field.  (This is
either specified by layout(location = ...) or assigned by the linker.)
Corresponding inputs/outputs will match up by location; if there's a
mismatch, we're allowed to have undefined behavior.

This may be less efficient - depending what locations were chosen, we
may have empty padding slots in the VUE.  But applications presumably
use small consecutive integers for locations, so it hopefully won't be
much worse in practice.

3% of Dota 2 Reborn shaders are hurt, but only by 2 instructions.
This seems like a small price to pay for avoiding recompiles.

Signed-off-by: Kenneth Graunke <[email protected]>
Reviewed-by: Chris Forbes <[email protected]>

1 files changed, 13 insertions, 1 deletions

diff --git a/src/mesa/drivers/dri/i965/brw_context.h b/src/mesa/drivers/dri/i965/brw_context.h
index 144d3e327d4..a7b612ad545 100644
--- a/src/mesa/drivers/dri/i965/brw_context.h
+++ b/src/mesa/drivers/dri/i965/brw_context.h
@@ -541,6 +541,17 @@ struct brw_vue_map {
    GLbitfield64 slots_valid;
 
    /**
+    * Is this VUE map for a separate shader pipeline?
+    *
+    * Separable programs (GL_ARB_separate_shader_objects) can be mixed and matched
+    * without the linker having a chance to dead code eliminate unused varyings.
+    *
+    * This means that we have to use a fixed slot layout, based on the output's
+    * location field, rather than assigning slots in a compact contiguous block.
+    */
+   bool separate;
+
+   /**
     * Map from gl_varying_slot value to VUE slot.  For gl_varying_slots that are
     * not stored in a slot (because they are not written, or because
     * additional processing is applied before storing them in the VUE), the
@@ -585,7 +596,8 @@ static inline GLuint brw_varying_to_offset(struct brw_vue_map *vue_map,
 
 void brw_compute_vue_map(const struct brw_device_info *devinfo,
                          struct brw_vue_map *vue_map,
-                         GLbitfield64 slots_valid);
+                         GLbitfield64 slots_valid,
+                         bool separate_shader);
 
 
 /**