gfx-rs · teoxoy · Dec 16, 2025 · Dec 13, 2025
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -165,6 +165,39 @@ event happens. Our new log policy is as follows:
 
 By @cwfitzgerald in [#8579](https://github.com/gfx-rs/wgpu/pull/8579).
 
+#### Push constants renamed immediates, API brought in line with spec.
+
+As the "immediate data" api is getting close to stabilization in the WebGPU specification,
+we're bringing our implementation in line with what the spec dictates.
+
+First, in the `PipelineLayoutDescriptor`, you now pass a unified size for all stages:
+
+```diff
+- push_constant_ranges: &[wgpu::PushConstantRange {
+-     stages: wgpu::ShaderStages::VERTEX_FRAGMENT,
+-     range: 0..12,
+- }]
++ immediate_size: 12,
+```
+
+Second, on the command encoder you no longer specify a shader stage, uploads apply
+to all shader stages that use immediate data.
+
+```diff
+- rpass.set_push_constants(wgpu::ShaderStages::FRAGMENT, 0, bytes);
++ rpass.set_immediates(0, bytes);
+```
+
+Finally, our implementation currently still zero-initializes the immediate data
+range you declared in the pipeline layout. This is not spec compliant and failing
+to populate immediate "slots" that are used in the shader will be a validation error
+in a future version. See [the proposal][immediate-data-spec] for details for determining
+which slots are populated in a given shader.
+
+By @cwfitzgerald in [#8724](https://github.com/gfx-rs/wgpu/pull/8724).
+
+[immediate-data-spec]: https://github.com/gpuweb/gpuweb/blob/main/proposals/immediate-data.md#immediate-slots
+
 #### `subgroup_{min,max}_size` renamed and moved from `Limits` -> `AdapterInfo`
 
 To bring our code in line with the WebGPU spec, we have moved information about subgroup size

diff --git a/benches/benches/wgpu-benchmark/computepass.rs b/benches/benches/wgpu-benchmark/computepass.rs
@@ -251,7 +251,7 @@ impl ComputepassState {
                 .create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
                     label: None,
                     bind_group_layouts: &[&bind_group_layout],
-                    immediates_ranges: &[],
+                    immediate_size: 0,
                 });
 
         let pipeline =
@@ -346,7 +346,7 @@ impl ComputepassState {
                     .create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
                         label: None,
                         bind_group_layouts: &[&bindless_bind_group_layout],
-                        immediates_ranges: &[],
+                        immediate_size: 0,
                     });
 
             let bindless_pipeline =

diff --git a/benches/benches/wgpu-benchmark/renderpass.rs b/benches/benches/wgpu-benchmark/renderpass.rs
@@ -151,7 +151,7 @@ impl RenderpassState {
                 .create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
                     label: None,
                     bind_group_layouts: &[&bind_group_layout],
-                    immediates_ranges: &[],
+                    immediate_size: 0,
                 });
 
         let mut vertex_buffers = Vec::with_capacity(vertex_buffer_count as usize);
@@ -287,7 +287,7 @@ impl RenderpassState {
                     .create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
                         label: None,
                         bind_group_layouts: &[&bindless_bind_group_layout],
-                        immediates_ranges: &[],
+                        immediate_size: 0,
                     });
 
             bindless_pipeline = Some(device_state.device.create_render_pipeline(

diff --git a/deno_webgpu/device.rs b/deno_webgpu/device.rs
@@ -380,7 +380,7 @@ impl GPUDevice {
     let wgpu_descriptor = wgpu_core::binding_model::PipelineLayoutDescriptor {
       label: crate::transform_label(descriptor.label.clone()),
       bind_group_layouts: Cow::Owned(bind_group_layouts),
-      immediates_ranges: Default::default(),
+      immediate_size: 0,
     };
 
     let (id, err) = self.instance.device_create_pipeline_layout(

diff --git a/examples/features/src/big_compute_buffers/mod.rs b/examples/features/src/big_compute_buffers/mod.rs
@@ -127,7 +127,7 @@ fn setup_pipeline(
     let pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
         label: Some("Compute Pipeline Layout"),
         bind_group_layouts: &[&bind_group_layout],
-        immediates_ranges: &[],
+        immediate_size: 0,
     });
 
     device.create_compute_pipeline(&wgpu::ComputePipelineDescriptor {

diff --git a/examples/features/src/boids/mod.rs b/examples/features/src/boids/mod.rs
@@ -107,7 +107,7 @@ impl crate::framework::Example for Example {
             device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
                 label: Some("compute"),
                 bind_group_layouts: &[&compute_bind_group_layout],
-                immediates_ranges: &[],
+                immediate_size: 0,
             });
 
         // create render pipeline with empty bind group layout
@@ -116,7 +116,7 @@ impl crate::framework::Example for Example {
             device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
                 label: Some("render"),
                 bind_group_layouts: &[],
-                immediates_ranges: &[],
+                immediate_size: 0,
             });
 
         let render_pipeline = device.create_render_pipeline(&wgpu::RenderPipelineDescriptor {

diff --git a/examples/features/src/bunnymark/mod.rs b/examples/features/src/bunnymark/mod.rs
@@ -201,7 +201,7 @@ impl crate::framework::Example for Example {
         let pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
             label: None,
             bind_group_layouts: &[&global_bind_group_layout, &local_bind_group_layout],
-            immediates_ranges: &[],
+            immediate_size: 0,
         });
 
         let pipeline = device.create_render_pipeline(&wgpu::RenderPipelineDescriptor {

diff --git a/examples/features/src/conservative_raster/mod.rs b/examples/features/src/conservative_raster/mod.rs
@@ -78,7 +78,7 @@ impl crate::framework::Example for Example {
             device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
                 label: None,
                 bind_group_layouts: &[],
-                immediates_ranges: &[],
+                immediate_size: 0,
             });
 
         let shader_triangle_and_lines =
@@ -195,7 +195,7 @@ impl crate::framework::Example for Example {
             let pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
                 label: None,
                 bind_group_layouts: &[&bind_group_layout],
-                immediates_ranges: &[],
+                immediate_size: 0,
             });
             let shader = device.create_shader_module(wgpu::include_wgsl!("upscale.wgsl"));
             (

diff --git a/examples/features/src/cube/mod.rs b/examples/features/src/cube/mod.rs
@@ -158,7 +158,7 @@ impl crate::framework::Example for Example {
         let pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
             label: None,
             bind_group_layouts: &[&bind_group_layout],
-            immediates_ranges: &[],
+            immediate_size: 0,
         });
 
         // Create the texture

diff --git a/examples/features/src/hello_synchronization/mod.rs b/examples/features/src/hello_synchronization/mod.rs
@@ -91,7 +91,7 @@ async fn execute(
     let pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
         label: None,
         bind_group_layouts: &[&bind_group_layout],
-        immediates_ranges: &[],
+        immediate_size: 0,
     });
     let patient_pipeline = device.create_compute_pipeline(&wgpu::ComputePipelineDescriptor {
         label: None,

diff --git a/examples/features/src/hello_triangle/mod.rs b/examples/features/src/hello_triangle/mod.rs
@@ -47,7 +47,7 @@ async fn run(event_loop: EventLoop<()>, window: Window) {
     let pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
         label: None,
         bind_group_layouts: &[],
-        immediates_ranges: &[],
+        immediate_size: 0,
     });
 
     let swapchain_capabilities = surface.get_capabilities(&adapter);

diff --git a/examples/features/src/hello_workgroups/mod.rs b/examples/features/src/hello_workgroups/mod.rs
@@ -100,7 +100,7 @@ async fn run() {
     let pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
         label: None,
         bind_group_layouts: &[&bind_group_layout],
-        immediates_ranges: &[],
+        immediate_size: 0,
     });
     let pipeline = device.create_compute_pipeline(&wgpu::ComputePipelineDescriptor {
         label: None,

diff --git a/examples/features/src/mesh_shader/mod.rs b/examples/features/src/mesh_shader/mod.rs
@@ -104,7 +104,7 @@ impl crate::framework::Example for Example {
         let pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
             label: None,
             bind_group_layouts: &[],
-            immediates_ranges: &[],
+            immediate_size: 0,
         });
         let pipeline = device.create_mesh_pipeline(&wgpu::MeshPipelineDescriptor {
             label: None,

diff --git a/examples/features/src/msaa_line/mod.rs b/examples/features/src/msaa_line/mod.rs
@@ -161,7 +161,7 @@ impl crate::framework::Example for Example {
         let pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
             label: None,
             bind_group_layouts: &[],
-            immediates_ranges: &[],
+            immediate_size: 0,
         });
 
         let multisampled_framebuffer =

diff --git a/examples/features/src/multiple_render_targets/mod.rs b/examples/features/src/multiple_render_targets/mod.rs
@@ -112,7 +112,7 @@ impl MultiTargetRenderer {
         let pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
             label: None,
             bind_group_layouts: &[&texture_bind_group_layout],
-            immediates_ranges: &[],
+            immediate_size: 0,
         });
 
         let sampler = device.create_sampler(&wgpu::SamplerDescriptor {
@@ -232,7 +232,7 @@ impl TargetRenderer {
         let pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
             label: None,
             bind_group_layouts: &[&texture_bind_group_layout],
-            immediates_ranges: &[],
+            immediate_size: 0,
         });
 
         let sampler = device.create_sampler(&wgpu::SamplerDescriptor {

diff --git a/examples/features/src/ray_shadows/mod.rs b/examples/features/src/ray_shadows/mod.rs
@@ -186,10 +186,7 @@ impl crate::framework::Example for Example {
         let pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
             label: None,
             bind_group_layouts: &[&bind_group_layout],
-            immediates_ranges: &[wgpu::ImmediateRange {
-                stages: wgpu::ShaderStages::FRAGMENT,
-                range: 0..12,
-            }],
+            immediate_size: 12,
         });
 
         let pipeline = device.create_render_pipeline(&wgpu::RenderPipelineDescriptor {
@@ -331,9 +328,9 @@ impl crate::framework::Example for Example {
 
             rpass.set_pipeline(&self.pipeline);
             rpass.set_bind_group(0, Some(&self.bind_group), &[]);
-            rpass.set_immediates(wgpu::ShaderStages::FRAGMENT, 0, &0.0_f32.to_ne_bytes());
-            rpass.set_immediates(wgpu::ShaderStages::FRAGMENT, 4, &cos.to_ne_bytes());
-            rpass.set_immediates(wgpu::ShaderStages::FRAGMENT, 8, &sin.to_ne_bytes());
+            rpass.set_immediates(0, &0.0_f32.to_ne_bytes());
+            rpass.set_immediates(4, &cos.to_ne_bytes());
+            rpass.set_immediates(8, &sin.to_ne_bytes());
             rpass.set_vertex_buffer(0, self.vertex_buf.slice(..));
             rpass.set_index_buffer(self.index_buf.slice(..), IndexFormat::Uint16);
             rpass.draw_indexed(0..12, 0, 0..1);

diff --git a/examples/features/src/ray_traced_triangle/mod.rs b/examples/features/src/ray_traced_triangle/mod.rs
@@ -269,7 +269,7 @@ impl crate::framework::Example for Example {
             device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
                 label: Some("pipeline layout for shader.wgsl"),
                 bind_group_layouts: &[&bgl],
-                immediates_ranges: &[],
+                immediate_size: 0,
             });
 
         let compute_pipeline = device.create_compute_pipeline(&wgpu::ComputePipelineDescriptor {
@@ -284,7 +284,7 @@ impl crate::framework::Example for Example {
         let blit_pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
             label: Some("pipeline layout for blit.wgsl"),
             bind_group_layouts: &[&blit_bgl],
-            immediates_ranges: &[],
+            immediate_size: 0,
         });
 
         let blit_pipeline = device.create_render_pipeline(&wgpu::RenderPipelineDescriptor {

diff --git a/examples/features/src/repeated_compute/mod.rs b/examples/features/src/repeated_compute/mod.rs
@@ -229,7 +229,7 @@ impl WgpuContext {
         let pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
             label: None,
             bind_group_layouts: &[&bind_group_layout],
-            immediates_ranges: &[],
+            immediate_size: 0,
         });
         let pipeline = device.create_compute_pipeline(&wgpu::ComputePipelineDescriptor {
             label: None,

diff --git a/examples/features/src/shadow/mod.rs b/examples/features/src/shadow/mod.rs
@@ -466,7 +466,7 @@ impl crate::framework::Example for Example {
             let pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
                 label: Some("shadow"),
                 bind_group_layouts: &[&bind_group_layout, &local_bind_group_layout],
-                immediates_ranges: &[],
+                immediate_size: 0,
             });
 
             let uniform_buf = device.create_buffer(&wgpu::BufferDescriptor {
@@ -582,7 +582,7 @@ impl crate::framework::Example for Example {
             let pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
                 label: Some("main"),
                 bind_group_layouts: &[&bind_group_layout, &local_bind_group_layout],
-                immediates_ranges: &[],
+                immediate_size: 0,
             });
 
             let mx_total = Self::generate_matrix(config.width as f32 / config.height as f32);

diff --git a/examples/features/src/skybox/mod.rs b/examples/features/src/skybox/mod.rs
@@ -187,7 +187,7 @@ impl crate::framework::Example for Example {
         let pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
             label: None,
             bind_group_layouts: &[&bind_group_layout],
-            immediates_ranges: &[],
+            immediate_size: 0,
         });
 
         // Create the render pipelines

diff --git a/examples/features/src/srgb_blend/mod.rs b/examples/features/src/srgb_blend/mod.rs
@@ -92,7 +92,7 @@ impl<const SRGB: bool> crate::framework::Example for Example<SRGB> {
         let pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
             label: None,
             bind_group_layouts: &[&bind_group_layout],
-            immediates_ranges: &[],
+            immediate_size: 0,
         });
 
         // Create bind group

diff --git a/examples/features/src/stencil_triangles/mod.rs b/examples/features/src/stencil_triangles/mod.rs
@@ -48,7 +48,7 @@ impl crate::framework::Example for Example {
         let pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
             label: None,
             bind_group_layouts: &[],
-            immediates_ranges: &[],
+            immediate_size: 0,
         });
 
         let shader = device.create_shader_module(wgpu::include_wgsl!("shader.wgsl"));

diff --git a/examples/features/src/storage_texture/mod.rs b/examples/features/src/storage_texture/mod.rs
@@ -90,7 +90,7 @@ async fn run(_path: Option<String>) {
     let pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
         label: None,
         bind_group_layouts: &[&bind_group_layout],
-        immediates_ranges: &[],
+        immediate_size: 0,
     });
     let pipeline = device.create_compute_pipeline(&wgpu::ComputePipelineDescriptor {
         label: None,

diff --git a/examples/features/src/texture_arrays/mod.rs b/examples/features/src/texture_arrays/mod.rs
@@ -328,7 +328,7 @@ impl crate::framework::Example for Example {
         let pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
             label: Some("main"),
             bind_group_layouts: &[&bind_group_layout, &uniform_bind_group_layout],
-            immediates_ranges: &[],
+            immediate_size: 0,
         });
 
         let index_format = wgpu::IndexFormat::Uint16;

diff --git a/examples/features/src/timestamp_queries/mod.rs b/examples/features/src/timestamp_queries/mod.rs
@@ -351,7 +351,7 @@ fn render_pass(
     let pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
         label: None,
         bind_group_layouts: &[],
-        immediates_ranges: &[],
+        immediate_size: 0,
     });
 
     let render_pipeline = device.create_render_pipeline(&wgpu::RenderPipelineDescriptor {

diff --git a/examples/features/src/uniform_values/mod.rs b/examples/features/src/uniform_values/mod.rs
@@ -162,7 +162,7 @@ impl WgpuContext {
             label: None,
             // (4)
             bind_group_layouts: &[&bind_group_layout],
-            immediates_ranges: &[],
+            immediate_size: 0,
         });
 
         let swapchain_capabilities = surface.get_capabilities(&adapter);

diff --git a/examples/features/src/water/mod.rs b/examples/features/src/water/mod.rs
@@ -428,14 +428,14 @@ impl crate::framework::Example for Example {
             device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
                 label: Some("water"),
                 bind_group_layouts: &[&water_bind_group_layout],
-                immediates_ranges: &[],
+                immediate_size: 0,
             });
 
         let terrain_pipeline_layout =
             device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
                 label: Some("terrain"),
                 bind_group_layouts: &[&terrain_bind_group_layout],
-                immediates_ranges: &[],
+                immediate_size: 0,
             });
 
         let water_uniform_buf = device.create_buffer(&wgpu::BufferDescriptor {

diff --git a/examples/standalone/01_hello_compute/src/main.rs b/examples/standalone/01_hello_compute/src/main.rs
@@ -168,7 +168,7 @@ fn main() {
     let pipeline_layout = device.create_pipeline_layout(&wgpu::PipelineLayoutDescriptor {
         label: None,
         bind_group_layouts: &[&bind_group_layout],
-        immediates_ranges: &[],
+        immediate_size: 0,
     });
 
     // The pipeline is the ready-to-go program state for the GPU. It contains the shader modules,

diff --git a/player/src/lib.rs b/player/src/lib.rs
@@ -202,7 +202,7 @@ impl Player {
                 let resolved_desc = wgc::binding_model::ResolvedPipelineLayoutDescriptor {
                     label: desc.label.clone(),
                     bind_group_layouts: Cow::from(&bind_group_layouts),
-                    immediates_ranges: Cow::Borrowed(&*desc.immediates_ranges),
+                    immediate_size: desc.immediate_size,
                 };
 
                 let pipeline_layout = device
@@ -1058,12 +1058,10 @@ impl Player {
             },
             C::SetScissor(rect) => C::SetScissor(rect),
             C::SetImmediate {
-                stages,
                 offset,
                 size_bytes,
                 values_offset,
             } => C::SetImmediate {
-                stages,
                 offset,
                 size_bytes,
                 values_offset,