[PATCH v2 0/1] MR97: vkd3d-shader/hlsl: Support all() intrinsic.

[Nikolay Sivov (＠nsivov)]

--
  v2: vkd3d-shader/hlsl: Support all() intrinsic.

https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/97

[PATCH v2 1/1] vkd3d-shader/hlsl: Support all() intrinsic.

[Nikolay Sivov]

From: Nikolay Sivov

---
 Makefile.am              |  1 +
 libs/vkd3d-shader/hlsl.y | 35 ++++++++++++++++++++++
 tests/all.shader_test    | 63 ++++++++++++++++++++++++++++++++++++++++
 3 files changed, 99 insertions(+)
 create mode 100644 tests/all.shader_test

diff --git a/Makefile.am b/Makefile.am
index f9199472..f7a149ee 100644
--- a/Makefile.am
+++ b/Makefile.am
@@ -43,6 +43,7 @@ vkd3d_cross_tests = \

vkd3d_shader_tests = \
    tests/abs.shader_test \
+	tests/all.shader_test \
    tests/arithmetic-float.shader_test \
    tests/arithmetic-float-uniform.shader_test \
    tests/arithmetic-int.shader_test \
diff --git a/libs/vkd3d-shader/hlsl.y b/libs/vkd3d-shader/hlsl.y
index 95c1c0fa..72da8ce4 100644
--- a/libs/vkd3d-shader/hlsl.y
+++ b/libs/vkd3d-shader/hlsl.y
@@ -2370,6 +2370,40 @@ static bool intrinsic_abs(struct hlsl_ctx *ctx,
     return !!add_unary_arithmetic_expr(ctx, params->instrs, HLSL_OP1_ABS, params->args[0], loc);
 }

+static bool intrinsic_all(struct hlsl_ctx *ctx,
+        const struct parse_initializer *params, const struct vkd3d_shader_location *loc)
+{
+    struct hlsl_ir_node *arg = params->args[0], *mul;
+    struct hlsl_type *arg_type = arg->data_type;
+    struct hlsl_ir_constant *one, *zero;
+    struct hlsl_ir_load *load;
+    unsigned int i, j;
+
+    if (!(one = hlsl_new_float_constant(ctx, 1.0f, loc)))
+        return false;
+    list_add_tail(params->instrs, &one->node.entry);
+
+    if (!(zero = hlsl_new_float_constant(ctx, 0.0f, loc)))
+        return false;
+    list_add_tail(params->instrs, &zero->node.entry);
+
+    mul = &one->node;
+
+    for (i = 0; i < arg_type->dimx; ++i)
+    {
+        for (j = 0; j < arg_type->dimy; ++j)
+        {
+            if (!(load = add_load_component(ctx, params->instrs, arg, j * arg->data_type->dimx + i, loc)))
+                return false;
+
+            if (!(mul = add_binary_arithmetic_expr(ctx, params->instrs, HLSL_OP2_MUL, &load->node, mul, loc)))
+                return false;
+        }
+    }
+
+    return !!add_binary_comparison_expr(ctx, params->instrs, HLSL_OP2_NEQUAL, mul, &zero->node, loc);
+}
+
 /* Find the type corresponding to the given source type, with the same
  * dimensions but a different base type. */
 static struct hlsl_type *convert_numeric_type(const struct hlsl_ctx *ctx,
@@ -2965,6 +2999,7 @@ intrinsic_functions[] =
 {
     /* Note: these entries should be kept in alphabetical order. */
     {"abs",                                 1, true,  intrinsic_abs},
+    {"all",                                 1, true,  intrinsic_all},
     {"asuint",                             -1, true,  intrinsic_asuint},
     {"clamp",                               3, true,  intrinsic_clamp},
     {"cos",                                 1, true,  intrinsic_cos},
diff --git a/tests/all.shader_test b/tests/all.shader_test
new file mode 100644
index 00000000..7bdb0dc8
--- /dev/null
+++ b/tests/all.shader_test
@@ -0,0 +1,63 @@
+[require]
+shader model >= 4.0
+
+[pixel shader]
+uniform float4 f;
+
+float4 main() : sv_target
+{
+    return all(f);
+}
+
+[test]
+uniform 0 float4 -1.1 1.6 1.3 0.5
+draw quad
+probe all rgba (1.0, 1.0, 1.0, 1.0)
+
+[test]
+uniform 0 float4 0.0 1.6 1.3 0.5
+draw quad
+probe all rgba (0.0, 0.0, 0.0, 0.0)
+
+[test]
+uniform 0 float4 1.0 0.0 1.3 0.5
+draw quad
+probe all rgba (0.0, 0.0, 0.0, 0.0)
+
+[pixel shader]
+uniform float f;
+
+float4 main() : sv_target
+{
+    return all(f);
+}
+
+[test]
+uniform 0 float4 1.0 0.0 0.0 0.0
+draw quad
+probe all rgba (1.0, 1.0, 1.0, 1.0)
+
+[test]
+uniform 0 float4 0.0 0.0 0.0 0.0
+draw quad
+probe all rgba (0.0, 0.0, 0.0, 0.0)
+
+[pixel shader]
+uniform float2x2 f;
+
+float4 main() : sv_target
+{
+    return all(f);
+}
+
+[test]
+uniform 0 float4 1.0 2.0 0.0 0.0
+uniform 4 float4 3.0 4.0 0.0 0.0
+draw quad
+probe all rgba (1.0, 1.0, 1.0, 1.0)
+
+[test]
+uniform 0 float4 1.0 2.0 0.0 0.0
+uniform 4 float4 0.0 4.0 0.0 0.0
+draw quad
+probe all rgba (0.0, 0.0, 0.0, 0.0)

[Nikolay Sivov (＠nsivov)]

On Wed Feb 15 20:46:17 2023 +0000, Zebediah Figura wrote:
> We don't actually need a synthetic variable for this, do we?

Good point.

[PATCH v2 0/1] MR97: vkd3d-shader/hlsl: Support all() intrinsic. - approved

[Zebediah Figura (＠zfigura)]

This merge request was approved by Zebediah Figura.

[Giovanni Mascellani (＠giomasce)]

Giovanni Mascellani (@giomasce) commented about libs/vkd3d-shader/hlsl.y:
> - if (!(one = hlsl_new_float_constant(ctx, 1.0f, loc)))
> -
>    return false;
> - list_add_tail(params->instrs, &one->node.entry);
> -
> - if (!(zero = hlsl_new_float_constant(ctx, 0.0f, loc)))
> -
>    return false;
> - list_add_tail(params->instrs, &zero->node.entry);
> -
> - mul = &one->node;
> -
> - for (i = 0; i < arg_type->dimx; ++i)
> - {
> -
>    for (j = 0; j < arg_type->dimy; ++j)
> -
>    {
> -
>        if (!(load = add_load_component(ctx, params->instrs, arg, j * arg->data_type->dimx + i, loc)))
> -
>            return false;

That's not wrong, but you could just iterate on a single index ranging from zero to `hlsl_type_component_count(arg_type)`.

[PATCH v2 0/1] MR97: vkd3d-shader/hlsl: Support all() intrinsic. - approved

[Giovanni Mascellani (＠giomasce)]

This merge request was approved by Giovanni Mascellani.

[Francisco Casas (＠fcasas)]

At first, I was suspicious of implementing `all()` using multiplication, but that seems to be what the native compiler does...

```
float4 main() : sv_target
{
    float4 a = {1, 1, 1, 0};
    float4 b = {1e20, 1e20, 1e20, 0};

return float4(all(a), all(b), -1, -1);
}
```

this actually returns `(0, 1, -1, -1)` both in native and using this patch, so it is correct in giving the incorrect result.

[PATCH v2 0/1] MR97: vkd3d-shader/hlsl: Support all() intrinsic. - approved

[Francisco Casas (＠fcasas)]

This merge request was approved by Francisco Casas.