Add immediate size validation cases for pipeline creation #4573

shaoboyan091 · 2026-01-21T02:31:38Z

Add cases to validate immediate size checks during both compute and render pipeline creation.

Issue: #

Requirements for PR author:

All missing test coverage is tracked with "TODO" or .unimplemented().
New helpers are /** documented */ and new helper files are found in helper_index.txt.
Test behaves as expected in a WebGPU implementation. (If not passing, explain above.)
Test have be tested with compatibility mode validation enabled and behave as expected. (If not passing, explain above.)

Requirements for reviewer sign-off:

Tests are properly located.
Test descriptions are accurate and complete.
Tests provide complete coverage (including validation control cases). Missing coverage MUST be covered by TODOs.
Tests avoid over-parameterization (see case count report).

When landing this PR, be sure to make any necessary issue status updates.

github-actions · 2026-01-21T02:32:18Z

Results for build job (at 2127eae):

+webgpu:api,validation,compute_pipeline:pipeline_creation_immediate_size_mismatch:* - 12 cases, 12 subcases (~1/case)
+webgpu:api,validation,render_pipeline,misc:pipeline_creation_immediate_size_mismatch:* - 18 cases, 18 subcases (~1/case)
-TOTAL: 285744 cases, 2344468 subcases
+TOTAL: 285774 cases, 2344498 subcases

Copilot

Pull request overview

This PR adds validation test cases for immediate data size checks during pipeline creation. The tests verify that pipeline creation fails when shaders use immediate data larger than the immediateSize specified in the pipeline layout or larger than maxImmediateSize when using automatic layout.

Changes:

Added helper function supportsImmediateData to check browser support for immediate data
Added pipeline_creation_immediate_size_mismatch test for compute pipelines
Added pipeline_creation_immediate_size_mismatch test for render pipelines

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 6 comments.

File	Description
src/common/util/util.ts	Added `supportsImmediateData` helper function to detect immediate data feature support
src/webgpu/api/validation/compute_pipeline.spec.ts	Added test to validate immediate size checks during compute pipeline creation
src/webgpu/api/validation/render_pipeline/misc.spec.ts	Added test to validate immediate size checks during render pipeline creation

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-01-21T02:36:44Z

src/webgpu/api/validation/render_pipeline/misc.spec.ts

+g.test('pipeline_creation_immediate_size_mismatch')
+  .desc(
+    `
+    Validate that creating a pipeline fails if the shader uses immediate data
+    larger than the immediateSize specified in the pipeline layout, or larger than
+    maxImmediateSize if layout is 'auto'.
+    Also validates that using less or equal size is allowed.
+    `
+  )
+  .params(u =>
+    u.combineWithParams([
+      { vertexSize: 16, fragmentSize: 16, layoutSize: 16 }, // Equal
+      { vertexSize: 12, fragmentSize: 12, layoutSize: 16 }, // Shader smaller
+      { vertexSize: 20, fragmentSize: 20, layoutSize: 16 }, // Shader larger (small diff)
+      { vertexSize: 32, fragmentSize: 32, layoutSize: 16 }, // Shader larger
+      { vertexSize: 'max+4', fragmentSize: 0, layoutSize: 'auto' }, // Vertex > Limit
+      { vertexSize: 0, fragmentSize: 'max+4', layoutSize: 'auto' }, // Fragment > Limit
+      { vertexSize: 'max', fragmentSize: 0, layoutSize: 'auto' }, // Vertex = Limit (Control)
+      { vertexSize: 0, fragmentSize: 'max', layoutSize: 'auto' }, // Fragment = Limit (Control)
+      { vertexSize: 'max', fragmentSize: 'max', layoutSize: 'auto' }, // Both at Limit (Control)
+    ] as const)
+  )
+  .fn(t => {
+    if (!supportsImmediateData(t.device)) {
+      t.skip('Immediate data not supported');
+    }
+    const { vertexSize, fragmentSize, layoutSize } = t.params;
+
+    const maxImmediateSize = t.device.limits.maxImmediateSize;
+    if (maxImmediateSize === undefined) {
+      t.skip('maxImmediateSize not supported');
+    }
+
+    const resolveSize = (sizeDescriptor: number | string) => {
+      if (typeof sizeDescriptor === 'number') return sizeDescriptor;
+      if (sizeDescriptor === 'max') return maxImmediateSize;
+      if (sizeDescriptor === 'max+4') return maxImmediateSize + 4;
+      return 0;
+    };
+
+    const resolvedVertexImmediateSize = resolveSize(vertexSize);
+    const resolvedFragmentImmediateSize = resolveSize(fragmentSize);
+    const varSize = Math.max(resolvedVertexImmediateSize, resolvedFragmentImmediateSize);
+
+    const code = `
+      var<immediate> data: array<u32, ${varSize / 4}>;
+      fn use_v() { ${resolvedVertexImmediateSize > 0 ? '_ = data[0];' : ''} }
+      fn use_f() { ${resolvedFragmentImmediateSize > 0 ? '_ = data[0];' : ''} }
+      @vertex fn main_vertex() -> @builtin(position) vec4<f32> { use_v(); return vec4<f32>(0.0, 0.0, 0.0, 1.0); }
+      @fragment fn main_fragment() -> @location(0) vec4<f32> { use_f(); return vec4<f32>(0.0, 1.0, 0.0, 1.0); }
+    `;
+
+    let layout: GPUPipelineLayout | 'auto';
+    let validSize: number;
+
+    if (layoutSize === 'auto') {
+      layout = 'auto';
+      validSize = maxImmediateSize;
+    } else {
+      layout = t.device.createPipelineLayout({
+        bindGroupLayouts: [],
+        immediateSize: layoutSize as number,
+      });
+      validSize = layoutSize as number;
+    }
+
+    const shouldError = varSize > validSize;
+
+    t.expectValidationError(() => {
+      t.device.createRenderPipeline({
+        layout,
+        vertex: {
+          module: t.device.createShaderModule({ code }),
+        },
+        fragment: {
+          module: t.device.createShaderModule({ code }),
+          targets: [{ format: 'rgba8unorm' }],
+        },
+      });
+    }, shouldError);
+  });


This test should validate both synchronous and asynchronous pipeline creation. Add an isAsync parameter to match the pattern used by other pipeline creation tests in this file. The params should include .combine('isAsync', [true, false]) and the test should use vtu.doCreateRenderPipelineTest helper with the isAsync parameter instead of directly calling createRenderPipeline.

Copilot · 2026-01-21T02:36:44Z

src/webgpu/api/validation/render_pipeline/misc.spec.ts

+          module: t.device.createShaderModule({ code }),
+        },
+        fragment: {
+          module: t.device.createShaderModule({ code }),


The shader defines entry points named main_vertex and main_fragment, but the pipeline descriptor doesn't specify entryPoint for either the vertex or fragment stages. According to the WebGPU specification, the default entry point name is "main", so the pipeline creation will fail with an entry point not found error instead of testing immediate data size validation.

Add entryPoint: 'main_vertex' to the vertex stage and entryPoint: 'main_fragment' to the fragment stage in the pipeline descriptor.

Suggested change

module: t.device.createShaderModule({ code }),

},

fragment: {

module: t.device.createShaderModule({ code }),

module: t.device.createShaderModule({ code }),

entryPoint: 'main_vertex',

},

fragment: {

module: t.device.createShaderModule({ code }),

entryPoint: 'main_fragment',

Copilot · 2026-01-21T02:36:45Z

src/webgpu/api/validation/compute_pipeline.spec.ts

+    ] as const)
+  )
+  .fn(t => {
+    if (!supportsImmediateData(t.device)) {


The supportsImmediateData function expects a GPU parameter but is being called with t.device which is a GPUDevice. This will cause a runtime error since GPUDevice does not have a wgslLanguageFeatures property.

The function should be called with the GPU object obtained via getGPU(t.rec) instead. Import getGPU from the navigator_gpu module and change the call to supportsImmediateData(getGPU(t.rec)).

Copilot · 2026-01-21T02:36:45Z

src/webgpu/api/validation/render_pipeline/misc.spec.ts

+    ] as const)
+  )
+  .fn(t => {
+    if (!supportsImmediateData(t.device)) {


The supportsImmediateData function expects a GPU parameter but is being called with t.device which is a GPUDevice. This will cause a runtime error since GPUDevice does not have a wgslLanguageFeatures property.

The function should be called with the GPU object obtained via getGPU(t.rec) instead. Import getGPU from the navigator_gpu module and change the call to supportsImmediateData(getGPU(t.rec)).

Copilot · 2026-01-21T02:36:45Z

src/webgpu/api/validation/compute_pipeline.spec.ts

+    const maxImmediateSize = t.device.limits.maxImmediateSize;
+
+    let actualLayout: GPUPipelineLayout | 'auto';
+    let validSize: number;
+
+    if (layoutSize === 'auto') {
+      actualLayout = 'auto';
+      // checked above
+      validSize = maxImmediateSize!;


The maxImmediateSize limit is accessed without checking if it's undefined. While there's a comment on line 849 saying "checked above", the only check above is the supportsImmediateData check which doesn't guarantee that maxImmediateSize is defined. Add a check similar to the render pipeline test: if (maxImmediateSize === undefined) { t.skip('maxImmediateSize not supported'); }

Copilot · 2026-01-21T02:36:45Z

src/webgpu/api/validation/compute_pipeline.spec.ts

+g.test('pipeline_creation_immediate_size_mismatch')
+  .desc(
+    `
+    Validate that creating a pipeline fails if the shader uses immediate data
+    larger than the immediateSize specified in the pipeline layout, or larger than
+    maxImmediateSize if layout is 'auto'.
+    Also validates that using less or equal size is allowed.
+    `
+  )
+  .params(u =>
+    u.combineWithParams([
+      { shaderSize: 16, layoutSize: 16 }, // Equal
+      { shaderSize: 12, layoutSize: 16 }, // Shader smaller
+      { shaderSize: 20, layoutSize: 16 }, // Shader larger (small diff)
+      { shaderSize: 32, layoutSize: 16 }, // Shader larger
+      { shaderSize: 'max', layoutSize: 'auto' }, // Shader equal to limit (auto layout)
+      { shaderSize: 'max+4', layoutSize: 'auto' }, // Shader larger than limit (auto layout)
+    ] as const)
+  )
+  .fn(t => {
+    if (!supportsImmediateData(t.device)) {
+      t.skip('Immediate data not supported');
+    }
+
+    const { shaderSize, layoutSize } = t.params;
+
+    const maxImmediateSize = t.device.limits.maxImmediateSize;
+
+    let actualLayout: GPUPipelineLayout | 'auto';
+    let validSize: number;
+
+    if (layoutSize === 'auto') {
+      actualLayout = 'auto';
+      // checked above
+      validSize = maxImmediateSize!;
+    } else {
+      actualLayout = t.device.createPipelineLayout({
+        bindGroupLayouts: [],
+        immediateSize: layoutSize as number,
+      });
+      validSize = layoutSize as number;
+    }
+
+    let actualShaderSize: number;
+    if (shaderSize === 'max') {
+      actualShaderSize = validSize;
+    } else if (shaderSize === 'max+4') {
+      actualShaderSize = validSize + 4;
+    } else {
+      actualShaderSize = shaderSize as number;
+    }
+
+    const code = `
+      var<immediate> data: array<u32, ${actualShaderSize / 4}>;
+      fn use() { _ = data[0]; }
+      @compute @workgroup_size(1) fn main_compute() { use(); }
+    `;
+
+    const shouldError = actualShaderSize > validSize;
+
+    t.expectValidationError(() => {
+      t.device.createComputePipeline({
+        layout: actualLayout,
+        compute: {
+          module: t.device.createShaderModule({ code }),
+          entryPoint: 'main_compute',
+        },
+      });
+    }, shouldError);
+  });


This test should validate both synchronous and asynchronous pipeline creation. Add an isAsync parameter to match the pattern used by other pipeline creation tests in this file. The params should include .combine('isAsync', [true, false]) and the test should use vtu.doCreateComputePipelineTest helper with the isAsync parameter instead of directly calling createComputePipeline.

Add cases to validate immediate size checks during both compute and render pipeline creation.

shaoboyan091 requested a review from Copilot January 21, 2026 02:31

Copilot started reviewing on behalf of shaoboyan091 January 21, 2026 02:32 View session

Copilot AI reviewed Jan 21, 2026

View reviewed changes

Add immediate size validation cases for pipeline creation

2127eae

Add cases to validate immediate size checks during both compute and render pipeline creation.

shaoboyan091 force-pushed the immediate-pipeline-creation branch from cf15532 to 2127eae Compare January 21, 2026 03:19

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add immediate size validation cases for pipeline creation #4573

Add immediate size validation cases for pipeline creation #4573

shaoboyan091 commented Jan 21, 2026

Uh oh!

github-actions bot commented Jan 21, 2026 •

edited

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Jan 21, 2026

Uh oh!

Copilot AI Jan 21, 2026

Uh oh!

Copilot AI Jan 21, 2026

Uh oh!

Copilot AI Jan 21, 2026

Uh oh!

Copilot AI Jan 21, 2026

Uh oh!

Copilot AI Jan 21, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Add immediate size validation cases for pipeline creation #4573

Are you sure you want to change the base?

Add immediate size validation cases for pipeline creation #4573

Conversation

shaoboyan091 commented Jan 21, 2026

Uh oh!

github-actions bot commented Jan 21, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Jan 21, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 21, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 21, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 21, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 21, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 21, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

github-actions bot commented Jan 21, 2026 •

edited

Loading