[Mesa-dev] [PATCH 01/15] i965/fs: Expose "urb_setup" as part of brw_wm_prog_data.

[Paul Berry]

At the moment, for Gen6+, the FS assumes that all varying inputs are
delivered to it in the order in which they appear in the
gl_program::InputsRead bitfield, and the SF/SBE setup code ensures
that they are delivered in this order.

When we add support for more than 64 varying components, this will no
longer always be possible, because the Gen6+ SF/SBE stage is only
capable of performing arbitrary reorderings of 16 varying slots.

To allow extra flexibility in the ordering of FS varyings, this patch
causes the FS to advertise exactly what ordering it expects.
---
 src/mesa/drivers/dri/i965/brw_context.h      |  7 +++++++
 src/mesa/drivers/dri/i965/brw_fs.cpp         | 10 +++++-----
 src/mesa/drivers/dri/i965/brw_fs.h           |  1 -
 src/mesa/drivers/dri/i965/brw_fs_visitor.cpp |  4 ++--
 4 files changed, 14 insertions(+), 8 deletions(-)

diff --git a/src/mesa/drivers/dri/i965/brw_context.h b/src/mesa/drivers/dri/i965/brw_context.h
index 939083b..41001d8 100644
--- a/src/mesa/drivers/dri/i965/brw_context.h
+++ b/src/mesa/drivers/dri/i965/brw_context.h
@@ -336,6 +336,13 @@ struct brw_wm_prog_data {
     */
    uint32_t barycentric_interp_modes;
 
+   /**
+    * Map from gl_varying_slot to the position within the FS setup data
+    * payload where the varying's attribute vertex deltas should be delivered.
+    * For varying slots that are not used by the FS, the value is -1.
+    */
+   int urb_setup[VARYING_SLOT_MAX];
+
    /* Pointers to tracked values (only valid once
     * _mesa_load_state_parameters has been called at runtime).
     *
diff --git a/src/mesa/drivers/dri/i965/brw_fs.cpp b/src/mesa/drivers/dri/i965/brw_fs.cpp
index 96cb2ee..9e7d203 100644
--- a/src/mesa/drivers/dri/i965/brw_fs.cpp
+++ b/src/mesa/drivers/dri/i965/brw_fs.cpp
@@ -1004,7 +1004,7 @@ fs_visitor::emit_general_interpolation(ir_variable *ir)
    int location = ir->location;
    for (unsigned int i = 0; i < array_elements; i++) {
       for (unsigned int j = 0; j < type->matrix_columns; j++) {
-	 if (urb_setup[location] == -1) {
+	 if (c->prog_data.urb_setup[location] == -1) {
 	    /* If there's no incoming setup data for this slot, don't
 	     * emit interpolation for it.
 	     */
@@ -1231,7 +1231,7 @@ void
 fs_visitor::calculate_urb_setup()
 {
    for (unsigned int i = 0; i < VARYING_SLOT_MAX; i++) {
-      urb_setup[i] = -1;
+      c->prog_data.urb_setup[i] = -1;
    }
 
    int urb_next = 0;
@@ -1239,7 +1239,7 @@ fs_visitor::calculate_urb_setup()
    if (brw->gen >= 6) {
       for (unsigned int i = 0; i < VARYING_SLOT_MAX; i++) {
 	 if (fp->Base.InputsRead & BITFIELD64_BIT(i)) {
-	    urb_setup[i] = urb_next++;
+	    c->prog_data.urb_setup[i] = urb_next++;
 	 }
       }
    } else {
@@ -1257,7 +1257,7 @@ fs_visitor::calculate_urb_setup()
 	     * incremented, mapped or not.
 	     */
 	    if (_mesa_varying_slot_in_fs((gl_varying_slot) i))
-	       urb_setup[i] = urb_next;
+	       c->prog_data.urb_setup[i] = urb_next;
             urb_next++;
 	 }
       }
@@ -1269,7 +1269,7 @@ fs_visitor::calculate_urb_setup()
        * See compile_sf_prog() for more info.
        */
       if (fp->Base.InputsRead & BITFIELD64_BIT(VARYING_SLOT_PNTC))
-         urb_setup[VARYING_SLOT_PNTC] = urb_next++;
+         c->prog_data.urb_setup[VARYING_SLOT_PNTC] = urb_next++;
    }
 
    /* Each attribute is 4 setup channels, each of which is half a reg. */
diff --git a/src/mesa/drivers/dri/i965/brw_fs.h b/src/mesa/drivers/dri/i965/brw_fs.h
index cb4ac3b..b77d4de 100644
--- a/src/mesa/drivers/dri/i965/brw_fs.h
+++ b/src/mesa/drivers/dri/i965/brw_fs.h
@@ -454,7 +454,6 @@ public:
    int first_non_payload_grf;
    /** Either BRW_MAX_GRF or GEN7_MRF_HACK_START */
    int max_grf;
-   int urb_setup[VARYING_SLOT_MAX];
 
    fs_reg *fp_temp_regs;
    fs_reg *fp_input_regs;
diff --git a/src/mesa/drivers/dri/i965/brw_fs_visitor.cpp b/src/mesa/drivers/dri/i965/brw_fs_visitor.cpp
index b049436..911b53e 100644
--- a/src/mesa/drivers/dri/i965/brw_fs_visitor.cpp
+++ b/src/mesa/drivers/dri/i965/brw_fs_visitor.cpp
@@ -2159,10 +2159,10 @@ fs_visitor::emit_dummy_fs()
 struct brw_reg
 fs_visitor::interp_reg(int location, int channel)
 {
-   int regnr = urb_setup[location] * 2 + channel / 2;
+   int regnr = c->prog_data.urb_setup[location] * 2 + channel / 2;
    int stride = (channel & 1) * 4;
 
-   assert(urb_setup[location] != -1);
+   assert(c->prog_data.urb_setup[location] != -1);
 
    return brw_vec1_grf(regnr, stride);
 }
-- 
1.8.4