/* Flushes undefined values to zero */
 
 static void
-optimise_nir(nir_shader *nir, unsigned quirks)
+optimise_nir(nir_shader *nir, unsigned quirks, bool is_blend)
 {
         bool progress;
         unsigned lower_flrp =
 
         NIR_PASS(progress, nir, midgard_nir_lower_algebraic_early);
 
+        if (!is_blend)
+                NIR_PASS(progress, nir, nir_fuse_io_16);
+
         do {
                 progress = false;
 
 
         /* Optimisation passes */
 
-        optimise_nir(nir, ctx->quirks);
+        optimise_nir(nir, ctx->quirks, is_blend);
 
         if (midgard_debug & MIDGARD_DBG_SHADERS) {
                 nir_print_shader(nir, stdout);
 
--- /dev/null
+/*
+ * Copyright (C) 2020 Collabora, Ltd.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors (Collabora):
+ *   Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
+ */
+
+/* Fuses f2f16 modifiers into loads */
+
+#include "compiler/nir/nir.h"
+#include "compiler/nir/nir_builder.h"
+#include "panfrost/util/pan_ir.h"
+
+bool nir_fuse_io_16(nir_shader *shader);
+
+static bool
+nir_src_is_f2fmp(nir_src *use)
+{
+   nir_instr *parent = use->parent_instr;
+
+   if (parent->type != nir_instr_type_alu)
+      return false;
+
+   nir_alu_instr *alu = nir_instr_as_alu(parent);
+   return (alu->op == nir_op_f2fmp);
+}
+
+bool
+nir_fuse_io_16(nir_shader *shader)
+{
+   bool progress = false;
+
+   nir_foreach_function(function, shader) {
+      if (!function->impl) continue;
+
+      nir_builder b;
+      nir_builder_init(&b, function->impl);
+
+      nir_foreach_block(block, function->impl) {
+         nir_foreach_instr_safe(instr, block) {
+            if (instr->type != nir_instr_type_intrinsic) continue;
+
+            nir_intrinsic_instr *intr = nir_instr_as_intrinsic(instr);
+
+            if (intr->intrinsic != nir_intrinsic_load_interpolated_input)
+                    continue;
+
+            if (nir_dest_bit_size(intr->dest) != 32)
+                    continue;
+
+            if (!intr->dest.is_ssa)
+               continue;
+
+            if (!list_is_empty(&intr->dest.ssa.if_uses))
+               return false;
+
+            bool valid = true;
+
+            nir_foreach_use(src, &intr->dest.ssa)
+               valid &= nir_src_is_f2fmp(src);
+
+            if (!valid)
+               continue;
+
+            intr->dest.ssa.bit_size = 16;
+            progress |= true;
+         }
+      }
+
+      nir_metadata_preserve(function->impl, nir_metadata_block_index | nir_metadata_dominance);
+
+   }
+
+   return progress;
+}