mesa: Add GL/GLSL plumbing for INTEL_fragment_shader_ordering

This extension provides new GLSL built-in function
beginFragmentShaderOrderingIntel() that guarantees
(taking wording of GL_INTEL_fragment_shader_ordering
extension) that any memory transactions issued by
shader invocations from previous primitives mapped to
same xy window coordinates (and same sample when
per-sample shading is active), complete and are visible
to the shader invocation that called
beginFragmentShaderOrderingINTEL().

One advantage of INTEL_fragment_shader_ordering over
ARB_fragment_shader_interlock is that it provides a
function that operates as a memory barrie (instead
of a defining a critcial section) that can be called
under arbitary control flow from any function (in
contrast the begin/end of ARB_fragment_shader_interlock
may only be called once, from main(), under no control
flow.

Signed-off-by: Kevin Rogovin <kevin.rogovin@intel.com>
Reviewed-by: Plamena Manolova <plamena.manolova@intel.com>

diff --git a/src/compiler/glsl/builtin_functions.cpp b/src/compiler/glsl/builtin_functions.cpp
index b6018806865054d8f7b6a1340e63605acf584664..5650365d1d549ccfd5085d8515be47456f107dde 100644 (file)
--- a/src/compiler/glsl/builtin_functions.cpp
+++ b/src/compiler/glsl/builtin_functions.cpp
@@ -525,6 +525,12 @@ supports_nv_fragment_shader_interlock(const _mesa_glsl_parse_state *state)
    return state->NV_fragment_shader_interlock_enable;
 }
 
+static bool
+supports_intel_fragment_shader_ordering(const _mesa_glsl_parse_state *state)
+{
+   return state->INTEL_fragment_shader_ordering_enable;
+}
+
 static bool
 shader_clock(const _mesa_glsl_parse_state *state)
 {
@@ -1305,6 +1311,11 @@ builtin_builder::create_intrinsics()
                    supports_arb_fragment_shader_interlock,
                    ir_intrinsic_end_invocation_interlock), NULL);
 
+   add_function("__intrinsic_begin_fragment_shader_ordering",
+                _invocation_interlock_intrinsic(
+                   supports_intel_fragment_shader_ordering,
+                   ir_intrinsic_begin_fragment_shader_ordering), NULL);
+
    add_function("__intrinsic_shader_clock",
                 _shader_clock_intrinsic(shader_clock,
                                         glsl_type::uvec2_type),
@@ -3419,6 +3430,12 @@ builtin_builder::create_builtins()
                    supports_nv_fragment_shader_interlock),
                 NULL);
 
+   add_function("beginFragmentShaderOrderingINTEL",
+                _invocation_interlock(
+                   "__intrinsic_begin_fragment_shader_ordering",
+                   supports_intel_fragment_shader_ordering),
+                NULL);
+
    add_function("anyInvocationARB",
                 _vote("__intrinsic_vote_any", vote),
                 NULL);

diff --git a/src/compiler/glsl/glsl_parser_extras.cpp b/src/compiler/glsl/glsl_parser_extras.cpp
index 0a7d0d78b14159fde0997d3e8ea0fa0164ea2ae2..21d41224440461a32b0c092f64c97190cb1c81b8 100644 (file)
--- a/src/compiler/glsl/glsl_parser_extras.cpp
+++ b/src/compiler/glsl/glsl_parser_extras.cpp
@@ -725,6 +725,7 @@ static const _mesa_glsl_extension _mesa_glsl_supported_extensions[] = {
    EXT_AEP(EXT_texture_buffer),
    EXT_AEP(EXT_texture_cube_map_array),
    EXT(INTEL_conservative_rasterization),
+   EXT(INTEL_fragment_shader_ordering),
    EXT(INTEL_shader_atomic_float_minmax),
    EXT(MESA_shader_integer_functions),
    EXT(NV_fragment_shader_interlock),

diff --git a/src/compiler/glsl/glsl_parser_extras.h b/src/compiler/glsl/glsl_parser_extras.h
index 2c8353214aa5ba8fe70b85709488ff32dc4a7b8c..e03b34d7d6c10dee7381d1632c4a67b84e51d3f7 100644 (file)
--- a/src/compiler/glsl/glsl_parser_extras.h
+++ b/src/compiler/glsl/glsl_parser_extras.h
@@ -812,6 +812,8 @@ struct _mesa_glsl_parse_state {
    bool EXT_texture_cube_map_array_warn;
    bool INTEL_conservative_rasterization_enable;
    bool INTEL_conservative_rasterization_warn;
+   bool INTEL_fragment_shader_ordering_enable;
+   bool INTEL_fragment_shader_ordering_warn;
    bool INTEL_shader_atomic_float_minmax_enable;
    bool INTEL_shader_atomic_float_minmax_warn;
    bool MESA_shader_integer_functions_enable;

diff --git a/src/compiler/glsl/glsl_to_nir.cpp b/src/compiler/glsl/glsl_to_nir.cpp
index a53000f47e9d8826ddc49574e7bdf63a36dced5b..efbb2317ac6e34f920290877aaea1d782d1334a5 100644 (file)
--- a/src/compiler/glsl/glsl_to_nir.cpp
+++ b/src/compiler/glsl/glsl_to_nir.cpp
@@ -742,6 +742,9 @@ nir_visitor::visit(ir_call *ir)
       case ir_intrinsic_end_invocation_interlock:
          op = nir_intrinsic_end_invocation_interlock;
          break;
+      case ir_intrinsic_begin_fragment_shader_ordering:
+         op = nir_intrinsic_begin_fragment_shader_ordering;
+         break;
       case ir_intrinsic_group_memory_barrier:
          op = nir_intrinsic_group_memory_barrier;
          break;
@@ -975,6 +978,9 @@ nir_visitor::visit(ir_call *ir)
       case nir_intrinsic_end_invocation_interlock:
          nir_builder_instr_insert(&b, &instr->instr);
          break;
+      case nir_intrinsic_begin_fragment_shader_ordering:
+         nir_builder_instr_insert(&b, &instr->instr);
+         break;
       case nir_intrinsic_store_ssbo: {
          exec_node *param = ir->actual_parameters.get_head();
          ir_rvalue *block = ((ir_instruction *)param)->as_rvalue();

diff --git a/src/compiler/glsl/ir.h b/src/compiler/glsl/ir.h
index d05d1998a50739f8937c36d0668f873ddd80a246..f478b29a6b5ecbaa53dbcab60acfacba4050ee45 100644 (file)
--- a/src/compiler/glsl/ir.h
+++ b/src/compiler/glsl/ir.h
@@ -1122,6 +1122,7 @@ enum ir_intrinsic_id {
    ir_intrinsic_memory_barrier_shared,
    ir_intrinsic_begin_invocation_interlock,
    ir_intrinsic_end_invocation_interlock,
+   ir_intrinsic_begin_fragment_shader_ordering,
 
    ir_intrinsic_vote_all,
    ir_intrinsic_vote_any,

diff --git a/src/compiler/nir/nir_intrinsics.py b/src/compiler/nir/nir_intrinsics.py
index 8c9bc3bbc9a46d7668103c7852d33a12ee5888cd..67d336fd536bc1bed9577bd7a6ff03183a2b9dff 100644 (file)
--- a/src/compiler/nir/nir_intrinsics.py
+++ b/src/compiler/nir/nir_intrinsics.py
@@ -191,6 +191,7 @@ barrier("memory_barrier_image")
 barrier("memory_barrier_shared")
 barrier("begin_invocation_interlock")
 barrier("end_invocation_interlock")
+barrier("begin_fragment_shader_ordering")
 
 # A conditional discard, with a single boolean source.
 intrinsic("discard_if", src_comp=[1])

diff --git a/src/mesa/main/extensions_table.h b/src/mesa/main/extensions_table.h
index 115af796bcc0618a96f9e53b39217d60ccd6f109..7072fa46ad860aada8113444260eef5c62ed7418 100644 (file)
--- a/src/mesa/main/extensions_table.h
+++ b/src/mesa/main/extensions_table.h
@@ -316,6 +316,7 @@ EXT(IBM_texture_mirrored_repeat             , dummy_true
 EXT(INGR_blend_func_separate                , EXT_blend_func_separate                , GLL,  x ,  x ,  x , 1999)
 
 EXT(INTEL_conservative_rasterization        , INTEL_conservative_rasterization       ,  x , GLC,  x ,  31, 2013)
+EXT(INTEL_fragment_shader_ordering          , INTEL_fragment_shader_ordering         , GLL, GLC,  x ,  x , 2013)
 EXT(INTEL_performance_query                 , INTEL_performance_query                , GLL, GLC,  x , ES2, 2013)
 EXT(INTEL_shader_atomic_float_minmax        , INTEL_shader_atomic_float_minmax       , GLL, GLC,  x ,  x , 2018)
 

diff --git a/src/mesa/main/mtypes.h b/src/mesa/main/mtypes.h
index 9d058cef6d91142dbaab8c1bc6aa953200bd0da4..f247b125e7e36a8d91955f4eace2757c740f0915 100644 (file)
--- a/src/mesa/main/mtypes.h
+++ b/src/mesa/main/mtypes.h
@@ -4271,6 +4271,7 @@ struct gl_extensions
    GLboolean ATI_fragment_shader;
    GLboolean GREMEDY_string_marker;
    GLboolean INTEL_conservative_rasterization;
+   GLboolean INTEL_fragment_shader_ordering;
    GLboolean INTEL_performance_query;
    GLboolean INTEL_shader_atomic_float_minmax;
    GLboolean KHR_blend_equation_advanced;