diff --git a/.github/workflows/ci.yaml b/.github/workflows/ci.yaml
index 68c6d2203e..be98878981 100644
--- a/.github/workflows/ci.yaml
+++ b/.github/workflows/ci.yaml
@@ -100,6 +100,8 @@ jobs:
         run: cargo fmt --all -- --check
       - name: Rustfmt tests
         run: rustfmt --check tests/ui/**/*.rs
+      - name: Check docs are valid
+        run: RUSTDOCFLAGS=-Dwarnings cargo doc
       - name: Clippy
         run: .github/workflows/clippy.sh
 
diff --git a/crates/spirv-std/src/arch/barrier.rs b/crates/spirv-std/src/arch/barrier.rs
index df255fab96..d93fd67373 100644
--- a/crates/spirv-std/src/arch/barrier.rs
+++ b/crates/spirv-std/src/arch/barrier.rs
@@ -23,10 +23,9 @@ use crate::memory::{Scope, Semantics};
 /// no such restriction starting with version 1.3.
 ///
 /// If used with the `TessellationControl` execution model, it also implicitly
-/// synchronizes the [`crate::storage_class::Output`] Storage Class: Writes to
-/// `Output` variables performed by any invocation executed prior to a
-/// [`control_barrier`] are visible to any other invocation proceeding beyond
-/// that [`control_barrier`].
+/// synchronizes the `output` storage class: Writes to `output` variables
+/// performed by any invocation executed prior to a [`control_barrier`] are
+/// visible to any other invocation proceeding beyond that [`control_barrier`].
 #[spirv_std_macros::gpu_only]
 #[doc(alias = "OpControlBarrier")]
 #[inline]
diff --git a/crates/spirv-std/src/arch/derivative.rs b/crates/spirv-std/src/arch/derivative.rs
index 0167e5ef61..6dede6bb5c 100644
--- a/crates/spirv-std/src/arch/derivative.rs
+++ b/crates/spirv-std/src/arch/derivative.rs
@@ -1,5 +1,3 @@
-//! The [`Derivative`] trait for getting derivatives and handling derivative
-//! operations in SPIR-V.
 use crate::float::Float;
 
 #[cfg(target_arch = "spirv")]
@@ -28,18 +26,17 @@ macro_rules! deriv_fn {
     };
 }
 
-/// Returns the partial derivative of `Self` with respect to the window's
-/// X coordinate. Returns the same result as either
-/// [`Self::ddx_fine`] or [`Self::ddx_coarse`], selection of which one is
-/// dependent on external factors.
+/// Returns the partial derivative of `component` with respect to the window's X
+/// coordinate. Returns the same result as either [`ddx_fine`] or
+/// [`ddx_coarse`], selection of which one is dependent on external factors.
 #[spirv_std_macros::vectorized]
 #[spirv_std_macros::gpu_only]
 pub fn ddx<F: Float>(component: F) -> F {
     deriv_fn!(component, OpDPdx, false)
 }
 
-/// Returns the partial derivative of `Self` with respect to the window's
-/// X coordinate. Uses local differencing based on the value of `Self` for
+/// Returns the partial derivative of `component` with respect to the window's X
+/// coordinate. Uses local differencing based on the value of `component` for
 /// the current fragment and its immediate neighbor(s).
 #[spirv_std_macros::vectorized]
 #[spirv_std_macros::gpu_only]
@@ -47,30 +44,29 @@ pub fn ddx_fine<F: Float>(component: F) -> F {
     deriv_fn!(component, OpDPdxFine, true)
 }
 
-/// Returns the partial derivative of `Self` with respect to the window's
-/// X coordinate. Uses local differencing based on the value of `Self` for
+/// Returns the partial derivative of `component` with respect to the window's X
+/// coordinate. Uses local differencing based on the value of `component` for
 /// the current fragment’s neighbors, and possibly, but not necessarily,
-/// includes the value of `Self` for the current fragment. That is, over a
-/// given area, the implementation can compute X derivatives in fewer
-/// unique locations than would be allowed by [`Self::ddx_fine`].
+/// includes the value of `component` for the current fragment. That is, over a
+/// given area, the implementation can compute X derivatives in fewer unique
+/// locations than would be allowed by [`ddx_fine`].
 #[spirv_std_macros::vectorized]
 #[spirv_std_macros::gpu_only]
 pub fn ddx_coarse<F: Float>(component: F) -> F {
     deriv_fn!(component, OpDPdxCoarse, true)
 }
 
-/// Returns the partial derivative of `Self` with respect to the window's
-/// Y coordinate. Returns the same result as either [`Self::ddy_fine`] or
-/// [`Self::ddy_coarse`], selection of which one is dependent on
-/// external factors.
+/// Returns the partial derivative of `component` with respect to the window's Y
+/// coordinate. Returns the same result as either [`ddy_fine`] or
+/// [`ddy_coarse`], selection of which one is dependent on external factors.
 #[spirv_std_macros::vectorized]
 #[spirv_std_macros::gpu_only]
 pub fn ddy<F: Float>(component: F) -> F {
     deriv_fn!(component, OpDPdy, false)
 }
 
-/// Returns the partial derivative of `Self` with respect to the window's
-/// Y coordinate. Uses local differencing based on the value of `Self` for
+/// Returns the partial derivative of `component` with respect to the window's Y
+/// coordinate. Uses local differencing based on the value of `component` for
 /// the current fragment and its immediate neighbor(s).
 #[spirv_std_macros::vectorized]
 #[spirv_std_macros::gpu_only]
@@ -78,36 +74,36 @@ pub fn ddy_fine<F: Float>(component: F) -> F {
     deriv_fn!(component, OpDPdyFine, true)
 }
 
-/// Returns the partial derivative of `Self` with respect to the window's
-/// Y coordinate. Uses local differencing based on the value of `Self` for
+/// Returns the partial derivative of `component` with respect to the window's Y
+/// coordinate. Uses local differencing based on the value of `component` for
 /// the current fragment’s neighbors, and possibly, but not necessarily,
-/// includes the value of `Self` for the current fragment. That is, over a
-/// given area, the implementation can compute Y derivatives in fewer
-/// unique locations than would be allowed by [`Derivative::ddy_fine`].
+/// includes the value of `component` for the current fragment. That is, over a
+/// given area, the implementation can compute Y derivatives in fewer unique
+/// locations than would be allowed by [`ddy_fine`].
 #[spirv_std_macros::vectorized]
 #[spirv_std_macros::gpu_only]
 pub fn ddy_coarse<F: Float>(component: F) -> F {
     deriv_fn!(component, OpDPdyCoarse, true)
 }
 
-/// Returns the sum of the absolute values of [`Self::ddx`] and
-/// [`Self::ddy`] as a single operation.
+/// Returns the sum of the absolute values of [`ddx`] and [`ddy`] as a single
+/// operation.
 #[spirv_std_macros::vectorized]
 #[spirv_std_macros::gpu_only]
 pub fn fwidth<F: Float>(component: F) -> F {
     deriv_fn!(component, OpFwidth, false)
 }
 
-/// Returns the sum of the absolute values of [`Self::ddx_fine`] and
-/// [`Self::ddy_fine`] as a single operation.
+/// Returns the sum of the absolute values of [`ddx_fine`] and [`ddy_fine`] as a
+/// single operation.
 #[spirv_std_macros::vectorized]
 #[spirv_std_macros::gpu_only]
 pub fn fwidth_fine<F: Float>(component: F) -> F {
     deriv_fn!(component, OpFwidthFine, true)
 }
 
-/// Returns the sum of the absolute values of [`Self::ddx_coarse`] and
-/// [`Self::ddy_coarse`] as a single operation.
+/// Returns the sum of the absolute values of [`ddx_coarse`] and [`ddy_coarse`]
+/// as a single operation.
 #[spirv_std_macros::vectorized]
 #[spirv_std_macros::gpu_only]
 pub fn fwidth_coarse<F: Float>(component: F) -> F {
diff --git a/crates/spirv-std/src/ray_tracing.rs b/crates/spirv-std/src/ray_tracing.rs
index 9ae62a2e1e..14ce58efaa 100644
--- a/crates/spirv-std/src/ray_tracing.rs
+++ b/crates/spirv-std/src/ray_tracing.rs
@@ -10,7 +10,7 @@ pub struct AccelerationStructure {
 }
 
 impl AccelerationStructure {
-    /// Converts a 64-bit integer into an [`AccelerationStructureKHR`].
+    /// Converts a 64-bit integer into an [`AccelerationStructure`].
     /// # Safety
     /// The 64-bit integer must point to a valid acceleration structure.
     #[spirv_std_macros::gpu_only]
@@ -31,7 +31,7 @@ impl AccelerationStructure {
         loop {}
     }
 
-    /// Converts a vector of two 32 bit integers into an [`AccelerationStructureKHR`].
+    /// Converts a vector of two 32 bit integers into an [`AccelerationStructure`].
     /// # Safety
     /// The combination must point to a valid acceleration structure.
     #[spirv_std_macros::gpu_only]