Skip to contentSkip to conversation

replace `allow(...)` lints with `expect(...)`

PATCH In review Onur Özkan 4 participants 17 messages Jul 1, 2025 – Jul 9, 2025

Subsystems (10)
lore
Share & tools
lore

Onur Özkan

author reviewer
#0 lore raw

v4 cover replace `allow(...)` lints with `expect(...)`

Changes in v4:
- The first patch from previous versions is now
broken down into multiple smaller patches.
- #![expect(dead_code)] is reverted from
rust/kernel/opp.rs as it failed on the
kernel test bot (see [1] for more).

Onur Özkan (6):
rust: switch to #[expect(...)] in core modules
rust: switch to #[expect(...)] in init and kunit
drivers: gpu: switch to #[expect(...)] in nova-core/regs.rs
rust: switch to #[expect(...)] in devres, driver and ioctl
rust: remove #[allow(clippy::unnecessary_cast)]
rust: remove #[allow(clippy::non_send_fields_in_send_ty)]

 drivers/gpu/nova-core/regs.rs       | 2 +-
 rust/kernel/alloc/allocator_test.rs | 2 +-
 rust/kernel/cpufreq.rs              | 1 -
 rust/kernel/devres.rs               | 2 +-
 rust/kernel/driver.rs               | 2 +-
 rust/kernel/drm/ioctl.rs            | 8 ++++----
 rust/kernel/error.rs                | 3 +--
 rust/kernel/init.rs                 | 6 +++---
 rust/kernel/kunit.rs                | 2 +-
 rust/kernel/types.rs                | 2 +-
 rust/macros/helpers.rs              | 2 +-
 11 files changed, 15 insertions(+), 17 deletions(-)

--
2.50.0

author reviewer
#1 lore raw

PATCH v4 1/6 rust: switch to `#[expect(...)]` in core modules

This makes it clear that the warning is expected not just ignored, so we don't end up having various unnecessary linting rules in the codebase.

Some parts of the codebase already use this approach, this patch just applies it more broadly.

No functional changes.

---
 rust/kernel/alloc/allocator_test.rs | 2 +-
 rust/kernel/error.rs                | 2 +-
 rust/kernel/types.rs                | 2 +-
 rust/macros/helpers.rs              | 2 +-
 4 files changed, 4 insertions(+), 4 deletions(-)
rust/kernel/alloc/allocator_test.rs +1 -1
index d19c06ef0498..844197d7194e 100644
@@ -7,7 +7,7 @@
 //! `Cmalloc` allocator within the `allocator_test` module and type alias all kernel allocators to
 //! `Cmalloc`. The `Cmalloc` allocator uses libc's `realloc()` function as allocator backend.
 
-#![allow(missing_docs)]
+#![expect(missing_docs)]
 
 use super::{flags::*, AllocError, Allocator, Flags};
 use core::alloc::Layout;
rust/kernel/error.rs +1 -1
index 3dee3139fcd4..a59613918d4c 100644
@@ -263,7 +263,7 @@ fn from(e: core::convert::Infallible) -> Error {
 /// [`samples/rust/rust_minimal.rs`]):
 ///
 /// ```
-/// # #[allow(clippy::single_match)]
+/// # #[expect(clippy::single_match)]
 /// fn example() -> Result {
 ///     let mut numbers = KVec::new();
 ///
rust/kernel/types.rs +1 -1
index 22985b6f6982..a5d5a4737a41 100644
@@ -600,5 +600,5 @@ pub enum Either<L, R> {
 /// constructed.
 ///
 /// [`NotThreadSafe`]: type@NotThreadSafe
-#[allow(non_upper_case_globals)]
+#[expect(non_upper_case_globals)]
 pub const NotThreadSafe: NotThreadSafe = PhantomData;
rust/macros/helpers.rs +1 -1
index e2602be402c1..6fd1462ff01a 100644
@@ -98,7 +98,7 @@ pub(crate) fn file() -> String {
     }
 
     #[cfg(CONFIG_RUSTC_HAS_SPAN_FILE)]
-    #[allow(clippy::incompatible_msrv)]
+    #[expect(clippy::incompatible_msrv)]
     {
         proc_macro::Span::call_site().file()
     }

--
2.50.0

reviewer
#2 lore raw

Re: [PATCH v4 1/6] rust: switch to `#[expect(...)]` in core modules

Quoted 17 lines · Onur Özkan

This makes it clear that the warning is expected not just ignored, so we don't end up having various unnecessary linting rules in the codebase.

Some parts of the codebase already use this approach, this patch just applies it more broadly.

No functional changes.

---

rust/kernel/alloc/allocator_test.rs +0 -0
index d19c06ef0498..844197d7194e 100644

rust/kernel/alloc/allocator_test.rs line 13

Quoted 8 lines
rust/kernel/alloc/allocator_test.rs +1 -1
@@ -7,7 +7,7 @@
 //! `Cmalloc` allocator within the `allocator_test` module and type alias all kernel allocators to
 //! `Cmalloc`. The `Cmalloc` allocator uses libc's `realloc()` function as allocator backend.
 
-#![allow(missing_docs)]
+#![expect(missing_docs)]
 
 use super::{flags::*, AllocError, Allocator, Flags};
 use core::alloc::Layout;

If you spin this again, would you mind adding a comment/reason about why there isn't documentation here?

This is preexisting but something to shift the possible meaning from "this module shouldn't have documentation" to "it's okay that we don't document unstable test interfaces" would be helpful.

Thanks,
Trevor

author reviewer
#3 lore raw

PATCH v4 2/6 rust: switch to `#[expect(...)]` in init and kunit

In reply to Onur Özkan

This makes it clear that the warning is expected not just ignored, so we don't end up having various unnecessary linting rules in the codebase.

Some parts of the codebase already use this approach, this patch just applies it more broadly.

No functional changes.

---
 rust/kernel/init.rs  | 6 +++---
 rust/kernel/kunit.rs | 2 +-
 2 files changed, 4 insertions(+), 4 deletions(-)
rust/kernel/init.rs +3 -3
index 8d228c237954..288b1c2a290d 100644
@@ -30,7 +30,7 @@
 //! ## General Examples
 //!
 //! ```rust,ignore
-//! # #![allow(clippy::disallowed_names)]
+//! # #![expect(clippy::disallowed_names)]
 //! use kernel::types::Opaque;
 //! use pin_init::pin_init_from_closure;
 //!
@@ -67,11 +67,11 @@
 //! ```
 //!
 //! ```rust,ignore
-//! # #![allow(unreachable_pub, clippy::disallowed_names)]
+//! # #![expect(unreachable_pub, clippy::disallowed_names)]
 //! use kernel::{prelude::*, types::Opaque};
 //! use core::{ptr::addr_of_mut, marker::PhantomPinned, pin::Pin};
 //! # mod bindings {
-//! #     #![allow(non_camel_case_types)]
+//! #     #![expect(non_camel_case_types)]
 //! #     pub struct foo;
 //! #     pub unsafe fn init_foo(_ptr: *mut foo) {}
 //! #     pub unsafe fn destroy_foo(_ptr: *mut foo) {}
rust/kernel/kunit.rs +1 -1
index 4b8cdcb21e77..91710a1d7b87 100644
@@ -280,7 +280,7 @@ macro_rules! kunit_unsafe_test_suite {
             static mut KUNIT_TEST_SUITE: ::kernel::bindings::kunit_suite =
                 ::kernel::bindings::kunit_suite {
                     name: KUNIT_TEST_SUITE_NAME,
-                    #[allow(unused_unsafe)]
+                    #[expect(unused_unsafe)]
                     // SAFETY: `$test_cases` is passed in by the user, and
                     // (as documented) must be valid for the lifetime of
                     // the suite (i.e., static).

--
2.50.0

maintainer
#4 lore raw

Re: [PATCH v4 2/6] rust: switch to `#[expect(...)]` in init and kunit

Quoted 7 lines · Onur Özkan

This makes it clear that the warning is expected not just ignored, so we don't end up having various unnecessary linting rules in the codebase.

Some parts of the codebase already use this approach, this patch just applies it more broadly.

Quoted 7 lines

No functional changes.

---
 rust/kernel/init.rs  | 6 +++---
 rust/kernel/kunit.rs | 2 +-
 2 files changed, 4 insertions(+), 4 deletions(-)

@Miguel are you going to pick this eventually, or do you think it should have a new version with the right splitting?

---
Cheers,
Benno

#5 lore raw

Re: [PATCH v4 2/6] rust: switch to `#[expect(...)]` in init and kunit

Quoted 3 lines · Benno Lossin

@Miguel are you going to pick this eventually, or do you think it should have a new version with the right splitting?

I guess I could take it since the series can only introduce build failures at worst.

However, I thought we were going to do the proper split (as independent patches, too) and then let each maintainer pick each on their own pace.

Cheers,
Miguel

maintainer
#6 lore raw

Re: [PATCH v4 2/6] rust: switch to `#[expect(...)]` in init and kunit

Quoted 11 lines · Miguel Ojeda
Quoted 3 lines · Benno Lossin

@Miguel are you going to pick this eventually, or do you think it should have a new version with the right splitting?

I guess I could take it since the series can only introduce build failures at worst.

However, I thought we were going to do the proper split (as independent patches, too) and then let each maintainer pick each on their own pace.

I also think it's better to split it correctly.

---
Cheers,
Benno

reviewer Reviewed-by on v4 2/6
#7 lore raw

Re: [PATCH v4 2/6] rust: switch to `#[expect(...)]` in init and kunit

In reply to Onur Özkan
Quoted 56 lines · Onur Özkan

This makes it clear that the warning is expected not just ignored, so we don't end up having various unnecessary linting rules in the codebase.

Some parts of the codebase already use this approach, this patch just applies it more broadly.

No functional changes.

---
 rust/kernel/init.rs  | 6 +++---
 rust/kernel/kunit.rs | 2 +-
 2 files changed, 4 insertions(+), 4 deletions(-)
rust/kernel/init.rs +3 -3
index 8d228c237954..288b1c2a290d 100644
@@ -30,7 +30,7 @@
 //! ## General Examples
 //!
 //! ```rust,ignore
-//! # #![allow(clippy::disallowed_names)]
+//! # #![expect(clippy::disallowed_names)]
 //! use kernel::types::Opaque;
 //! use pin_init::pin_init_from_closure;
 //!
@@ -67,11 +67,11 @@
 //! ```
 //!
 //! ```rust,ignore
-//! # #![allow(unreachable_pub, clippy::disallowed_names)]
+//! # #![expect(unreachable_pub, clippy::disallowed_names)]
 //! use kernel::{prelude::*, types::Opaque};
 //! use core::{ptr::addr_of_mut, marker::PhantomPinned, pin::Pin};
 //! # mod bindings {
-//! #     #![allow(non_camel_case_types)]
+//! #     #![expect(non_camel_case_types)]
 //! #     pub struct foo;
 //! #     pub unsafe fn init_foo(_ptr: *mut foo) {}
 //! #     pub unsafe fn destroy_foo(_ptr: *mut foo) {}
rust/kernel/kunit.rs +1 -2
index 4b8cdcb21e77..91710a1d7b87 100644
@@ -280,7 +280,7 @@ macro_rules! kunit_unsafe_test_suite {
             static mut KUNIT_TEST_SUITE: ::kernel::bindings::kunit_suite =
                 ::kernel::bindings::kunit_suite {
                     name: KUNIT_TEST_SUITE_NAME,
-                    #[allow(unused_unsafe)]
+                    #[expect(unused_unsafe)]
                     // SAFETY: `$test_cases` is passed in by the user, and
                     // (as documented) must be valid for the lifetime of
                     // the suite (i.e., static).
--
Quoted 1 line

2.50.0

Understood that the files may wind up split here, but the changes look good to me.

author reviewer
#8 lore raw

PATCH v4 3/6 drivers: gpu: switch to `#[expect(...)]` in nova-core/regs.rs

In reply to Onur Özkan

This makes it clear that the warning is expected not just ignored, so we don't end up having various unnecessary linting rules in the codebase.

Some parts of the codebase already use this approach, this patch just applies it more broadly.

No functional changes.

---
 drivers/gpu/nova-core/regs.rs | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/nova-core/regs.rs +1 -1
index 5a1273230306..87e5963f1ebb 100644
@@ -2,7 +2,7 @@
 
 // Required to retain the original register names used by OpenRM, which are all capital snake case
 // but are mapped to types.
-#![allow(non_camel_case_types)]
+#![expect(non_camel_case_types)]
 
 #[macro_use]
 mod macros;

--
2.50.0

reviewer Reviewed-by on v4 3/6
#9 lore raw

Re: [PATCH v4 3/6] drivers: gpu: switch to `#[expect(...)]` in nova-core/regs.rs

Quoted 28 lines · Onur Özkan

This makes it clear that the warning is expected not just ignored, so we don't end up having various unnecessary linting rules in the codebase.

Some parts of the codebase already use this approach, this patch just applies it more broadly.

No functional changes.

---
 drivers/gpu/nova-core/regs.rs | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/nova-core/regs.rs +1 -2
index 5a1273230306..87e5963f1ebb 100644
@@ -2,7 +2,7 @@
 
 // Required to retain the original register names used by OpenRM, which are all capital snake case
 // but are mapped to types.
-#![allow(non_camel_case_types)]
+#![expect(non_camel_case_types)]
 
 #[macro_use]
 mod macros;
--
Quoted 1 line

2.50.0

author reviewer
#10 lore raw

PATCH v4 4/6 rust: switch to `#[expect(...)]` in devres, driver and ioctl

In reply to Onur Özkan

This makes it clear that the warning is expected not just ignored, so we don't end up having various unnecessary linting rules in the codebase.

Some parts of the codebase already use this approach, this patch just applies it more broadly.

No functional changes.

---
 rust/kernel/devres.rs    | 2 +-
 rust/kernel/driver.rs    | 2 +-
 rust/kernel/drm/ioctl.rs | 8 ++++----
 3 files changed, 6 insertions(+), 6 deletions(-)
rust/kernel/devres.rs +1 -1
index 57502534d985..0e9510cf4625 100644
@@ -157,7 +157,7 @@ fn remove_action(this: &Arc<Self>) -> bool {
         success
     }
 
-    #[allow(clippy::missing_safety_doc)]
+    #[expect(clippy::missing_safety_doc)]
     unsafe extern "C" fn devres_callback(ptr: *mut kernel::ffi::c_void) {
         let ptr = ptr as *mut DevresInner<T>;
         // Devres owned this memory; now that we received the callback, drop the `Arc` and hence the
rust/kernel/driver.rs +1 -1
index ec9166cedfa7..fc7bd65b01f1 100644
@@ -168,7 +168,7 @@ fn of_id_info(dev: &device::Device) -> Option<&'static Self::IdInfo> {
     }
 
     #[cfg(not(CONFIG_OF))]
-    #[allow(missing_docs)]
+    #[expect(missing_docs)]
     fn of_id_info(_dev: &device::Device) -> Option<&'static Self::IdInfo> {
         None
     }
rust/kernel/drm/ioctl.rs +4 -4
index 445639404fb7..3ae8d2d8263f 100644
@@ -9,28 +9,28 @@
 const BASE: u32 = uapi::DRM_IOCTL_BASE as u32;
 
 /// Construct a DRM ioctl number with no argument.
-#[allow(non_snake_case)]
+#[expect(non_snake_case)]
 #[inline(always)]
 pub const fn IO(nr: u32) -> u32 {
     ioctl::_IO(BASE, nr)
 }
 
 /// Construct a DRM ioctl number with a read-only argument.
-#[allow(non_snake_case)]
+#[expect(non_snake_case)]
 #[inline(always)]
 pub const fn IOR<T>(nr: u32) -> u32 {
     ioctl::_IOR::<T>(BASE, nr)
 }
 
 /// Construct a DRM ioctl number with a write-only argument.
-#[allow(non_snake_case)]
+#[expect(non_snake_case)]
 #[inline(always)]
 pub const fn IOW<T>(nr: u32) -> u32 {
     ioctl::_IOW::<T>(BASE, nr)
 }
 
 /// Construct a DRM ioctl number with a read-write argument.
-#[allow(non_snake_case)]
+#[expect(non_snake_case)]
 #[inline(always)]
 pub const fn IOWR<T>(nr: u32) -> u32 {
     ioctl::_IOWR::<T>(BASE, nr)

--
2.50.0

author reviewer
#11 lore raw

PATCH v4 5/6 rust: remove `#[allow(clippy::unnecessary_cast)]`

In reply to Onur Özkan

This isn't needed anymore since kernel::ffi::c_int type is always i32 which differs from err (isize).

---
 rust/kernel/error.rs | 1 -
 1 file changed, 1 deletion(-)
rust/kernel/error.rs +0 -1
index a59613918d4c..05c6e71c0afb 100644
@@ -413,7 +413,6 @@ pub fn from_err_ptr<T>(ptr: *mut T) -> Result<*mut T> {
         // SAFETY: The FFI function does not deref the pointer.
         let err = unsafe { bindings::PTR_ERR(const_ptr) };
 
-        #[allow(clippy::unnecessary_cast)]
         // CAST: If `IS_ERR()` returns `true`,
         // then `PTR_ERR()` is guaranteed to return a
         // negative value greater-or-equal to `-bindings::MAX_ERRNO`,

--
2.50.0

reviewer Reviewed-by on v4 5/6
#12 lore raw

Re: [PATCH v4 5/6] rust: remove `#[allow(clippy::unnecessary_cast)]`

Quoted 21 lines · Onur Özkan

This isn't needed anymore since kernel::ffi::c_int type is always i32 which differs from err (isize).

---
 rust/kernel/error.rs | 1 -
 1 file changed, 1 deletion(-)
rust/kernel/error.rs +0 -2
index a59613918d4c..05c6e71c0afb 100644
@@ -413,7 +413,6 @@ pub fn from_err_ptr<T>(ptr: *mut T) -> Result<*mut T> {
         // SAFETY: The FFI function does not deref the pointer.
         let err = unsafe { bindings::PTR_ERR(const_ptr) };
 
-        #[allow(clippy::unnecessary_cast)]
         // CAST: If `IS_ERR()` returns `true`,
         // then `PTR_ERR()` is guaranteed to return a
         // negative value greater-or-equal to `-bindings::MAX_ERRNO`,
--
Quoted 1 line

2.50.0

author reviewer
#13 lore raw

PATCH v4 6/6 rust: remove `#[allow(clippy::non_send_fields_in_send_ty)]`

In reply to Onur Özkan

Just like how we did in [1], we no longer need this lint as it's no longer part of Clippy's suspicious lint group.

---
 rust/kernel/cpufreq.rs | 1 -
 1 file changed, 1 deletion(-)
rust/kernel/cpufreq.rs +0 -1
index 11b03e9d7e89..97de9b0573da 100644
@@ -907,7 +907,6 @@ fn register_em(_policy: &mut Policy) {
 /// or CPUs, so it is safe to share it.
 unsafe impl<T: Driver> Sync for Registration<T> {}
 
-#[allow(clippy::non_send_fields_in_send_ty)]
 /// SAFETY: Registration with and unregistration from the cpufreq subsystem can happen from any
 /// thread.
 unsafe impl<T: Driver> Send for Registration<T> {}

--
2.50.0

#14 lore raw

Re: [PATCH v4 6/6] rust: remove `#[allow(clippy::non_send_fields_in_send_ty)]`

Quoted 5 lines · Onur Özkan

Just like how we did in [1], we no longer need this lint as it's no longer part of Clippy's suspicious lint group.

Normally this format is used:

... did in commit 5e7c9b84ad08 ("rust: sync: remove unneeded
#[allow(clippy::non_send_fields_in_send_ty)]"), we no longer ...

Maintainers may want to fix this when they apply the patch, so I wouldn't send a new version just for this (but if you do, then please reword it).

Cheers,
Miguel

#15 lore raw
In reply to Onur Özkan
Quoted 8 lines · Onur Özkan

Onur Özkan (6):
rust: switch to #[expect(...)] in core modules
rust: switch to #[expect(...)] in init and kunit
drivers: gpu: switch to #[expect(...)] in nova-core/regs.rs
rust: switch to #[expect(...)] in devres, driver and ioctl
rust: remove #[allow(clippy::unnecessary_cast)]
rust: remove #[allow(clippy::non_send_fields_in_send_ty)]

Thanks for splitting this, it is very much appreciated.

I think you decided on a split based on similarity (?). In the kernel, different maintainers/trees take care of different areas/files (the MAINTAINERS file describes this mapping). For instance, in one of them, KUnit and init is mixed, which are maintained by different people.

That is why the split I suggested was one that allows maintainers to pick them independently. With this one, maintainers still have to coordinate, which isn't a big deal, since the changes are small and straightforward, but the purpose of the split was to avoid that.

Cheers,
Miguel

author reviewer
#16 lore raw
Quoted 10 lines · On Tue, 1 Jul 2025 10:43:10 +0200 Miguel Ojeda…
Quoted 8 lines · Onur Özkan

Onur Özkan (6):
rust: switch to #[expect(...)] in core modules
rust: switch to #[expect(...)] in init and kunit
drivers: gpu: switch to #[expect(...)] in nova-core/regs.rs
rust: switch to #[expect(...)] in devres, driver and ioctl
rust: remove #[allow(clippy::unnecessary_cast)]
rust: remove #[allow(clippy::non_send_fields_in_send_ty)]

Quoted 7 lines

Thanks for splitting this, it is very much appreciated.

I think you decided on a split based on similarity (?). In the kernel, different maintainers/trees take care of different areas/files (the MAINTAINERS file describes this mapping). For instance, in one of them, KUnit and init is mixed, which are maintained by different people.

I didn't know I could use MAINTAINERS file to organize my changes.

Quoted 4 lines

That is why the split I suggested was one that allows maintainers to pick them independently. With this one, maintainers still have to coordinate, which isn't a big deal, since the changes are small and straightforward, but the purpose of the split was to avoid that.

I see, I will make sure to use MAINTAINERS file as a ref next time, thank you. :)

Regards,
Onur