vkd3d-shader/hlsl: Implement the InterlockedXor() intrinsic.

Approved-by: Henri Verbeet (@hverbeet)
Approved-by: Elizabeth Figura (@zfigura)
Merge-Request: https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/1330

libs/vkd3d-shader/hlsl.c

@@ -3513,6 +3513,7 @@ static void dump_ir_interlocked(struct vkd3d_string_buffer *buffer, const struct
         [HLSL_INTERLOCKED_MAX] = "max",
         [HLSL_INTERLOCKED_MIN] = "min",
         [HLSL_INTERLOCKED_OR] = "or",
+        [HLSL_INTERLOCKED_XOR] = "xor",
     };
 
     VKD3D_ASSERT(interlocked->op < ARRAY_SIZE(op_names));

libs/vkd3d-shader/hlsl.h

@@ -960,6 +960,7 @@ enum hlsl_interlocked_op
     HLSL_INTERLOCKED_MAX,
     HLSL_INTERLOCKED_MIN,
     HLSL_INTERLOCKED_OR,
+    HLSL_INTERLOCKED_XOR,
 };
 
 /* Represents an interlocked operation.

libs/vkd3d-shader/hlsl.y

@@ -5548,6 +5548,12 @@ static bool intrinsic_InterlockedOr(struct hlsl_ctx *ctx,
     return intrinsic_interlocked(ctx, HLSL_INTERLOCKED_OR, params, loc, "InterlockedOr");
 }
 
+static bool intrinsic_InterlockedXor(struct hlsl_ctx *ctx,
+        const struct parse_initializer *params, const struct vkd3d_shader_location *loc)
+{
+    return intrinsic_interlocked(ctx, HLSL_INTERLOCKED_XOR, params, loc, "InterlockedXor");
+}
+
 static const struct intrinsic_function
 {
     const char *name;
@@ -5569,6 +5575,7 @@ intrinsic_functions[] =
     {"InterlockedMax",                     -1, true,  intrinsic_InterlockedMax},
     {"InterlockedMin",                     -1, true,  intrinsic_InterlockedMin},
     {"InterlockedOr",                      -1, true,  intrinsic_InterlockedOr},
+    {"InterlockedXor",                     -1, true,  intrinsic_InterlockedXor},
     {"abs",                                 1, true,  intrinsic_abs},
     {"acos",                                1, true,  intrinsic_acos},
     {"all",                                 1, true,  intrinsic_all},

libs/vkd3d-shader/hlsl_codegen.c

@@ -10006,6 +10006,7 @@ static bool sm4_generate_vsir_instr_interlocked(struct hlsl_ctx *ctx,
         [HLSL_INTERLOCKED_MAX] = VKD3DSIH_ATOMIC_UMAX,
         [HLSL_INTERLOCKED_MIN] = VKD3DSIH_ATOMIC_UMIN,
         [HLSL_INTERLOCKED_OR] = VKD3DSIH_ATOMIC_OR,
+        [HLSL_INTERLOCKED_XOR] = VKD3DSIH_ATOMIC_XOR,
     };
 
     static const enum vkd3d_shader_opcode imm_opcodes[] =
@@ -10017,6 +10018,7 @@ static bool sm4_generate_vsir_instr_interlocked(struct hlsl_ctx *ctx,
         [HLSL_INTERLOCKED_MAX] = VKD3DSIH_IMM_ATOMIC_UMAX,
         [HLSL_INTERLOCKED_MIN] = VKD3DSIH_IMM_ATOMIC_UMIN,
         [HLSL_INTERLOCKED_OR] = VKD3DSIH_IMM_ATOMIC_OR,
+        [HLSL_INTERLOCKED_XOR] = VKD3DSIH_IMM_ATOMIC_XOR,
     };
 
     struct hlsl_ir_node *cmp_value = interlocked->cmp_value.node, *value = interlocked->value.node;

libs/vkd3d-shader/tpf.c

@@ -4034,6 +4034,7 @@ static void tpf_handle_instruction(struct tpf_compiler *tpf, const struct vkd3d_
         case VKD3DSIH_ATOMIC_UMAX:
         case VKD3DSIH_ATOMIC_UMIN:
         case VKD3DSIH_ATOMIC_OR:
+        case VKD3DSIH_ATOMIC_XOR:
         case VKD3DSIH_AND:
         case VKD3DSIH_BREAK:
         case VKD3DSIH_CASE:
@@ -4085,6 +4086,7 @@ static void tpf_handle_instruction(struct tpf_compiler *tpf, const struct vkd3d_
         case VKD3DSIH_IMM_ATOMIC_UMAX:
         case VKD3DSIH_IMM_ATOMIC_UMIN:
         case VKD3DSIH_IMM_ATOMIC_OR:
+        case VKD3DSIH_IMM_ATOMIC_XOR:
         case VKD3DSIH_IMUL:
         case VKD3DSIH_INE:
         case VKD3DSIH_INEG:

tests/hlsl/uav-atomics.shader_test

@@ -8,7 +8,7 @@ size (buffer, 11)
 
 0xd 5 6 0x10 4 4 7 2 0 5 0
 
-[compute shader fail(sm<5) todo(sm>=5)]
+[compute shader fail(sm<5)]
 RWBuffer<uint> u : register(u1);
 
 uniform uint4 v;
@@ -32,7 +32,7 @@ void main()
 
 [test]
 uniform 0 uint4 3 5 0 0
-todo(sm<6) dispatch 1 1 1
+todo(glsl | msl) dispatch 1 1 1
 probe uav 1 (0) rui (1)
 probe uav 1 (1) rui (3)
 probe uav 1 (2) rui (15)
@@ -46,7 +46,7 @@ probe uav 1 (9) rui (3)
 probe uav 1 (10) rui (1)
 
 uniform 0 uint4 1 2 0 0
-todo(sm<6) dispatch 2 1 1
+todo(glsl | msl) dispatch 2 1 1
 probe uav 1 (0) rui (1)
 probe uav 1 (1) rui (3)
 probe uav 1 (2) rui (21)
@@ -96,7 +96,7 @@ size (2d, 11, 1)
 
 0xd 5 6 0x10 4 4 7 2 0 5 0
 
-[compute shader fail(sm<5) todo(sm>=5)]
+[compute shader fail(sm<5)]
 RWTexture2D<uint> u : register(u1);
 
 uniform uint4 v;
@@ -120,7 +120,7 @@ void main()
 
 [test]
 uniform 0 uint4 3 5 0 0
-todo(sm<6) dispatch 1 1 1
+todo(glsl | msl) dispatch 1 1 1
 probe uav 1 (0) rui (1)
 probe uav 1 (1) rui (3)
 probe uav 1 (2) rui (15)