i965/fs: add emit_shader_float_controls_execution_mode() and aux functions

We need this function to emit code that setups the control register
later with the defined execution mode for the shader. Therefore, we
emit it as the first instruction.

v2:
- Fix bug in setting the default mode mask in brw_rnd_mode_from_nir().
- Fix support for rounding modes in brw_rnd_mode_from_nir().

v3:
- Updated to renamed shader info member and enum values (Andres).

v4:
- Add actual emission as first instruction of emit_nir_code (Caio).

Signed-off-by: Samuel Iglesias Gonsálvez <[email protected]>
Signed-off-by: Andres Gomez <[email protected]>
Reviewed-by: Caio Marcelo de Oliveira Filho <[email protected]>

3 files changed, 61 insertions, 0 deletions

diff --git a/src/intel/compiler/brw_fs.h b/src/intel/compiler/brw_fs.h
index 6eef13bb317..0701c274adc 100644
--- a/src/intel/compiler/brw_fs.h
+++ b/src/intel/compiler/brw_fs.h
@@ -187,6 +187,7 @@ public:
    void emit_discard_jump();
    void emit_fsign(const class brw::fs_builder &, const nir_alu_instr *instr,
                    fs_reg result, fs_reg *op, unsigned fsign_src);
+   void emit_shader_float_controls_execution_mode();
    bool opt_peephole_sel();
    bool opt_peephole_csel();
    bool opt_peephole_predicated_break();
diff --git a/src/intel/compiler/brw_fs_nir.cpp b/src/intel/compiler/brw_fs_nir.cpp
index 5ed35ff78c0..42b85da32c8 100644
--- a/src/intel/compiler/brw_fs_nir.cpp
+++ b/src/intel/compiler/brw_fs_nir.cpp
@@ -34,6 +34,8 @@ using namespace brw;
 void
 fs_visitor::emit_nir_code()
 {
+   emit_shader_float_controls_execution_mode();
+
    /* emit the arrays used for inputs and outputs - load/store intrinsics will
     * be converted to reads/writes of these arrays
     */
diff --git a/src/intel/compiler/brw_fs_visitor.cpp b/src/intel/compiler/brw_fs_visitor.cpp
index 49eb9c00f1b..d723e2818e8 100644
--- a/src/intel/compiler/brw_fs_visitor.cpp
+++ b/src/intel/compiler/brw_fs_visitor.cpp
@@ -200,6 +200,64 @@ fs_visitor::emit_interpolation_setup_gen4()
    abld.emit(SHADER_OPCODE_RCP, this->pixel_w, wpos_w);
 }
 
+static unsigned
+brw_rnd_mode_from_nir(unsigned mode, unsigned *mask)
+{
+   unsigned brw_mode = 0;
+   *mask = 0;
+
+   if ((FLOAT_CONTROLS_ROUNDING_MODE_RTZ_FP16 |
+        FLOAT_CONTROLS_ROUNDING_MODE_RTZ_FP32 |
+        FLOAT_CONTROLS_ROUNDING_MODE_RTZ_FP64) &
+       mode) {
+      brw_mode |= BRW_RND_MODE_RTZ << BRW_CR0_RND_MODE_SHIFT;
+      *mask |= BRW_CR0_RND_MODE_MASK;
+   }
+   if ((FLOAT_CONTROLS_ROUNDING_MODE_RTE_FP16 |
+        FLOAT_CONTROLS_ROUNDING_MODE_RTE_FP32 |
+        FLOAT_CONTROLS_ROUNDING_MODE_RTE_FP64) &
+       mode) {
+      brw_mode |= BRW_RND_MODE_RTNE << BRW_CR0_RND_MODE_SHIFT;
+      *mask |= BRW_CR0_RND_MODE_MASK;
+   }
+   if (mode & FLOAT_CONTROLS_DENORM_PRESERVE_FP16) {
+      brw_mode |= BRW_CR0_FP16_DENORM_PRESERVE;
+      *mask |= BRW_CR0_FP16_DENORM_PRESERVE;
+   }
+   if (mode & FLOAT_CONTROLS_DENORM_PRESERVE_FP32) {
+      brw_mode |= BRW_CR0_FP32_DENORM_PRESERVE;
+      *mask |= BRW_CR0_FP32_DENORM_PRESERVE;
+   }
+   if (mode & FLOAT_CONTROLS_DENORM_PRESERVE_FP64) {
+      brw_mode |= BRW_CR0_FP64_DENORM_PRESERVE;
+      *mask |= BRW_CR0_FP64_DENORM_PRESERVE;
+   }
+   if (mode & FLOAT_CONTROLS_DENORM_FLUSH_TO_ZERO_FP16)
+      *mask |= BRW_CR0_FP16_DENORM_PRESERVE;
+   if (mode & FLOAT_CONTROLS_DENORM_FLUSH_TO_ZERO_FP32)
+      *mask |= BRW_CR0_FP32_DENORM_PRESERVE;
+   if (mode & FLOAT_CONTROLS_DENORM_FLUSH_TO_ZERO_FP64)
+      *mask |= BRW_CR0_FP64_DENORM_PRESERVE;
+   if (mode == FLOAT_CONTROLS_DEFAULT_FLOAT_CONTROL_MODE)
+      *mask |= BRW_CR0_FP_MODE_MASK;
+
+   return brw_mode;
+}
+
+void
+fs_visitor::emit_shader_float_controls_execution_mode()
+{
+   unsigned execution_mode = this->nir->info.float_controls_execution_mode;
+   if (execution_mode == FLOAT_CONTROLS_DEFAULT_FLOAT_CONTROL_MODE)
+      return;
+
+   fs_builder abld = bld.annotate("shader floats control execution mode");
+   unsigned mask = 0;
+   unsigned mode = brw_rnd_mode_from_nir(execution_mode, &mask);
+   abld.emit(SHADER_OPCODE_FLOAT_CONTROL_MODE, bld.null_reg_ud(),
+             brw_imm_d(mode), brw_imm_d(mask));
+}
+
 /** Emits the interpolation for the varying inputs. */
 void
 fs_visitor::emit_interpolation_setup_gen6()