spirq 0.4.4

Light weight SPIR-V query utility for graphics.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223

use glsl_to_spirv_macros::*;
use glsl_to_spirv_macros_impl::*;
use std::collections::HashSet;
use super::*;

macro_rules! gen_entries(
    ($stage_fn:ident, $glsl:expr) => {{
        static SPV: &'static [u8] = $stage_fn!($glsl);
        SpirvBinary::from(SPV).reflect().unwrap()
    }}
);
macro_rules! gen_one_entry(
    ($stage_fn:ident, $glsl:expr) => {{
        let entries = gen_entries!($stage_fn, $glsl);
        assert_eq!(entries.len(), 1, "expected 1 entry point, found {}", entries.len());
        entries[0].clone()
    }}
);

#[test]
fn test_basic_shader_stage() {
    macro_rules! x(
        ($stage_fn:ident, $exec_model:ident) => {{
            let entry = gen_one_entry!($stage_fn, "#version 450 core\nvoid main() {}");
            assert_eq!(entry.exec_model, ExecutionModel::$exec_model, "shader stage execution model mismatched");
            assert!(entry.spec.spec_consts().next().is_none(), "unexpected specialization");
            assert!(entry.inputs().next().is_none(), "unexpected input");
            assert!(entry.outputs().next().is_none(), "unexpected output");
            assert!(entry.descs().next().is_none(), "unexpected descriptor binding");
        }}
    );
    x!(glsl_vs, Vertex);
    x!(glsl_fs, Fragment);
}
// For maximal compatibility interface varriables like the output of vertex
// buffer is not forced empty, because it requires prior knowledge from the
// users, that this shader is a vertex shader. The reflection algorithm should
// be able to extract metadata for ANY type of shaders, including those haven't
// been created yet.
#[test]
fn test_vs_input_loc() {
    let entry = gen_one_entry!(glsl_vs, r#"
        #version 450 core
        layout(location=0, component=0)
        in uint a;
        layout(location=0, component=1)
        in vec3 b;
        layout(location=1)
        in vec4 c;
        layout(location=3)
        in uvec4 d;
        layout(location=4, component=2)
        in uvec2 e;
        void main() { gl_Position = vec4(a, b) + c + vec4(d) + vec4(e,0,0); }
    "#);
    let locations = entry.inputs()
        .map(|x| x.location)
        .collect::<HashSet<_>>();
    assert!(locations.contains(&InterfaceLocation(0, 0)));
    assert!(locations.contains(&InterfaceLocation(0, 1)));
    assert!(locations.contains(&InterfaceLocation(1, 0)));
    assert!(locations.contains(&InterfaceLocation(3, 0)));
    assert!(locations.contains(&InterfaceLocation(4, 2)));
    assert!(!locations.contains(&InterfaceLocation(0, 2)));
    assert!(!locations.contains(&InterfaceLocation(1, 1)));
    // Test for consistency.
    for input in entry.inputs() {
        let resolved = entry.resolve_input(entry.get_input_name(input.location).unwrap()).unwrap();
        assert_eq!(input, resolved);
        assert_eq!(resolved.ty, entry.get_input(input.location).unwrap());
    }
}
#[test]
fn test_fs_output_loc() {
    // Note that fragment shader location-sharing outputs must have the same
    // type.
    let entry = gen_one_entry!(glsl_fs, r#"
        #version 450 core
        layout(location=0, component=0)
        out float a;
        layout(location=0, component=1)
        out vec3 b;
        layout(location=1)
        out vec4 c;
        layout(location=3)
        out uvec4 d;
        layout(location=4, component=2)
        out uvec2 e;
        void main() { a = 0; b = vec3(0,0,0); c = vec4(0,0,0,0); d = uvec4(0,0,0,0); e = uvec2(0,0); }
    "#);
    let locations = entry.outputs()
        .map(|x| x.location)
        .collect::<HashSet<_>>();
    assert!(locations.contains(&InterfaceLocation(0, 0)));
    assert!(locations.contains(&InterfaceLocation(0, 1)));
    assert!(locations.contains(&InterfaceLocation(1, 0)));
    assert!(locations.contains(&InterfaceLocation(3, 0)));
    assert!(locations.contains(&InterfaceLocation(4, 2)));
    assert!(!locations.contains(&InterfaceLocation(0, 2)));
    assert!(!locations.contains(&InterfaceLocation(1, 1)));
    // Test for consistency.
    for output in entry.outputs() {
        let resolved = entry.resolve_output(entry.get_output_name(output.location).unwrap()).unwrap();
        assert_eq!(output, resolved);
        assert_eq!(resolved.ty, entry.get_output(output.location).unwrap());
    }
}
#[test]
fn test_spec_consts() {
    let entry = gen_one_entry!(glsl_gs, r#"
        #version 450 core
        layout(points) in;
        layout(points, max_vertices=1) out;
        layout(constant_id=233) const float a = 0;
        layout(constant_id=234) const float b = 0;
        layout(constant_id=235) const float c = 0;
        void main() { gl_Position = vec4(a,b,c,0); EmitVertex(); EndPrimitive(); }
    "#);
    let spec_ids = entry.spec.spec_consts()
        .map(|x| x.spec_id)
        .collect::<HashSet<_>>();
    assert!(spec_ids.contains(&233));
    assert!(spec_ids.contains(&234));
    assert!(spec_ids.contains(&235));
    assert!(!spec_ids.contains(&0));
    assert!(!spec_ids.contains(&1));
    assert!(!spec_ids.contains(&236));
    // Test for consistency.
    for spec_const in entry.spec.spec_consts() {
        let resolved = entry.spec.resolve_spec_const(
            entry.spec.get_spec_const_name(spec_const.spec_id).unwrap()
        ).unwrap();
        assert_eq!(spec_const, resolved);
        assert_eq!(spec_const.ty, entry.spec.get_spec_const(spec_const.spec_id).unwrap());
    }
}
#[test]
fn test_desc_tys() {
    let entry = gen_one_entry!(glsl_fs, r#"
        #version 450 core
        layout(std140, set=0, binding=0)
        uniform A {
            uint a;
        } aa;
        layout(std430, binding=1)
        buffer B {
            float b;
        } bb;
        layout(set=1, binding=0)
        uniform sampler2D c;
        layout(rgba32f, set=3, binding=4) readonly
        uniform image2D d;
        layout(input_attachment_index=0, set=1, binding=3)
        uniform subpassInput e;
        layout(std430, binding=3)
        buffer F {
            vec4 f;
        } ff;
        void main() {
            bb.b = 0.0;
            ff.f = vec4(aa.a,0,0,0) + bb.b * texture(c, vec2(0,0)) + subpassLoad(e) + imageLoad(d, ivec2(0,0));
        }
    "#);
    let desc_binds = entry.descs()
        .map(|x| x.desc_bind)
        .collect::<HashSet<_>>();
    assert!(desc_binds.contains(&DescriptorBinding(0, 0)));
    assert!(desc_binds.contains(&DescriptorBinding(0, 1)));
    assert!(desc_binds.contains(&DescriptorBinding(1, 0)));
    assert!(desc_binds.contains(&DescriptorBinding(3, 4)));
    assert!(desc_binds.contains(&DescriptorBinding(1, 3)));
    assert!(desc_binds.contains(&DescriptorBinding(0, 3)));
    assert!(!desc_binds.contains(&DescriptorBinding(0, 2)));
    // Test for consistency.
    for desc in entry.descs() {
        let resolved = entry.resolve_desc(entry.get_desc_name(desc.desc_bind).unwrap()).unwrap();
        assert_eq!(desc, resolved);
        assert_eq!(resolved.desc_ty, entry.get_desc(desc.desc_bind).unwrap());
        if desc.desc_bind == DescriptorBinding(0, 3) {
            assert_eq!(entry.get_desc_access(desc.desc_bind).unwrap(), AccessType::WriteOnly);
        } else if desc.desc_bind == DescriptorBinding(0, 1) {
            assert_eq!(entry.get_desc_access(desc.desc_bind).unwrap(), AccessType::ReadWrite);
        } else {
            assert_eq!(entry.get_desc_access(desc.desc_bind).unwrap(), AccessType::ReadOnly);
        }
    }
}
#[test]
fn test_push_consts() {
    let entry = gen_one_entry!(glsl_vs, r#"
        #version 450 core
        layout(push_constant)
        uniform A {
            float a;
        } aa;
        void main() { gl_Position = vec4(aa.a,0,0,0); }
    "#);
    let resolved = entry.resolve_push_const(
        entry.get_push_const_name().unwrap()
    ).unwrap();
    assert_eq!(entry.get_push_const().unwrap(), resolved.ty);
    assert_eq!(entry.get_push_const().unwrap().resolve(""), resolved.member_var_res);
}
#[test]
fn test_implicit_sampled_img() {
    use shaderc::{CompileOptions, SourceLanguage, ShaderKind, Compiler};
    // Currently only shaderc is outputting binding-sharing image and sampler.
    let src = r#"
        [[vk::bind(0, 0)]]
        Texture2D img;
        [[vk::bind(0, 0)]]
        SamplerState samp;
        float4 main() : SV_POSITION { return img.Sample(samp, float2(0,0)); }
    "#;
    let mut opt = CompileOptions::new().unwrap();
    opt.set_source_language(SourceLanguage::HLSL);
    let mut compiler = Compiler::new().unwrap();
    let out = compiler.compile_into_spirv(src, ShaderKind::Vertex, "<inline>", "main", Some(&opt))
        .unwrap();
    let spv: Vec<u32> = out.as_binary().into();
    SpirvBinary::from(spv).reflect().unwrap();
}
// TODO: (penguinliong) Comprehensive type testing.