Skip to contentSkip to conversation

rust: acpi: add `acpi::DeviceId` abstraction

PATCH In review Igor Korotin 3 participants 8 messages May 30, 2025 – Jun 3, 2025

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 rust: acpi: add `acpi::DeviceId` abstraction

acpi::DeviceId is an abstraction around struct acpi_device_id.

This is used by subsequent patches, in particular the i2c driver abstractions, to create ACPI device ID tables.

---
 MAINTAINERS         |  1 +
 rust/kernel/acpi.rs | 63 +++++++++++++++++++++++++++++++++++++++++++++
 rust/kernel/lib.rs  |  1 +
 3 files changed, 65 insertions(+)
 create mode 100644 rust/kernel/acpi.rs
MAINTAINERS +1 -0
index b659fb27ab63..5f8dfae08454 100644
@@ -302,6 +302,7 @@ F:	include/linux/acpi.h
 F:	include/linux/fwnode.h
 F:	include/linux/fw_table.h
 F:	lib/fw_table.c
+F:	rust/kernel/acpi.rs
 F:	tools/power/acpi/
 
 ACPI APEI
rust/kernel/acpi.rs new file +63 -0

63 lines not on this page · raw message

rust/kernel/lib.rs +1 -0
index 15c5f72976cd..05f1d3870bf7 100644
@@ -78,6 +78,7 @@
 #[cfg(CONFIG_NET)]
 pub mod net;
 pub mod of;
+pub mod acpi;
 pub mod page;
 #[cfg(CONFIG_I2C)]
 pub mod i2c;
Signature · 1 line
2.43.0
#1 lore raw
Quoted 5 lines · Igor Korotin

acpi::DeviceId is an abstraction around struct acpi_device_id.

This is used by subsequent patches, in particular the i2c driver abstractions, to create ACPI device ID tables.

Quoted 7 lines
---
 MAINTAINERS         |  1 +
 rust/kernel/acpi.rs | 63 +++++++++++++++++++++++++++++++++++++++++++++
 rust/kernel/lib.rs  |  1 +
 3 files changed, 65 insertions(+)
 create mode 100644 rust/kernel/acpi.rs

I'm not seeing any "subsequent patches" here. Is this part of a patch series that didn't all get sent out properly?

thanks,

greg k-h

Quoted 87 lines
MAINTAINERS +1 -0
index b659fb27ab63..5f8dfae08454 100644
@@ -302,6 +302,7 @@ F:	include/linux/acpi.h
 F:	include/linux/fwnode.h
 F:	include/linux/fw_table.h
 F:	lib/fw_table.c
+F:	rust/kernel/acpi.rs
 F:	tools/power/acpi/
 
 ACPI APEI
rust/kernel/acpi.rs new file +63 -0
new file mode 100644
index 000000000000..bbd38910736c
@@ -0,0 +1,63 @@
+// SPDX-License-Identifier: GPL-2.0
+
+//! Advanced Configuration and Power Interface abstractions.
+
+use crate::{bindings, device_id::RawDeviceId, prelude::*};
+
+/// IdTable type for ACPI drivers.
+pub type IdTable<T> = &'static dyn kernel::device_id::IdTable<DeviceId, T>;
+
+/// An ACPI device id.
+#[repr(transparent)]
+#[derive(Clone, Copy)]
+pub struct DeviceId(bindings::acpi_device_id);
+
+// SAFETY:
+// * `DeviceId` is a `#[repr(transparent)` wrapper of `struct acpi_device_id` and does not add
+//   additional invariants, so it's safe to transmute to `RawType`.
+// * `DRIVER_DATA_OFFSET` is the offset to the `data` field.
+unsafe impl RawDeviceId for DeviceId {
+    type RawType = bindings::acpi_device_id;
+
+    const DRIVER_DATA_OFFSET: usize = core::mem::offset_of!(bindings::acpi_device_id, driver_data);
+
+    fn index(&self) -> usize {
+        self.0.driver_data as _
+    }
+}
+
+impl DeviceId {
+    const ACPI_ID_LEN: usize = 16;
+
+    /// Create a new device id from an ACPI 'id' string.
+    pub const fn new(id: &'static CStr) -> Self {
+        assert!(id.len() <= Self::ACPI_ID_LEN, "ID exceeds 16 bytes");
+        let src = id.as_bytes_with_nul();
+        // Replace with `bindings::acpi_device_id::default()` once stabilized for `const`.
+        // SAFETY: FFI type is valid to be zero-initialized.
+        let mut acpi: bindings::acpi_device_id = unsafe { core::mem::zeroed() };
+        // TODO: Use `clone_from_slice` once the corresponding types do match.
+        let mut i = 0;
+        while i < src.len() {
+            acpi.id[i] = src[i] as _;
+            i += 1;
+        }
+
+        Self(acpi)
+    }
+}
+
+/// Create an ACPI `IdTable` with an "alias" for modpost.
+#[macro_export]
+macro_rules! acpi_device_table {
+    ($table_name:ident, $module_table_name:ident, $id_info_type: ty, $table_data: expr) => {
+        const $table_name: $crate::device_id::IdArray<
+            $crate::acpi::DeviceId,
+            $id_info_type,
+            { $table_data.len() },
+        > = $crate::device_id::IdArray::new($table_data);
+
+        $crate::module_device_table!("acpi", $module_table_name, $table_name);
+    };
+}
+
rust/kernel/lib.rs +0 -0
index 15c5f72976cd..05f1d3870bf7 100644

rust/kernel/lib.rs line 81

Quoted 4 lines
rust/kernel/lib.rs +1 -0
@@ -78,6 +78,7 @@
 #[cfg(CONFIG_NET)]
 pub mod net;
 pub mod of;
+pub mod acpi;

Aren't these supposed to be sorted?

thanks,

greg k-h

#2 lore raw
In reply to Igor Korotin
Quoted 6 lines · Igor Korotin

acpi::DeviceId is an abstraction around struct acpi_device_id.

This is used by subsequent patches, in particular the i2c driver abstractions, to create ACPI device ID tables.

As Greg mentioned it would be nice to see the subsequent patches.

Also, for this to be useful you need to implement the generic glue code in rust/kernel/driver.rs in the generic Adapter trait.

You also may want to connect it to the platform bus abstraction, just like the OF bindings are, since I assume that's the bus you care about?

Quoted 53 lines
---
 MAINTAINERS         |  1 +
 rust/kernel/acpi.rs | 63 +++++++++++++++++++++++++++++++++++++++++++++
 rust/kernel/lib.rs  |  1 +
 3 files changed, 65 insertions(+)
 create mode 100644 rust/kernel/acpi.rs
MAINTAINERS +1 -0
index b659fb27ab63..5f8dfae08454 100644
@@ -302,6 +302,7 @@ F:	include/linux/acpi.h
 F:	include/linux/fwnode.h
 F:	include/linux/fw_table.h
 F:	lib/fw_table.c
+F:	rust/kernel/acpi.rs
 F:	tools/power/acpi/
 
 ACPI APEI
rust/kernel/acpi.rs new file +28 -0
new file mode 100644
index 000000000000..bbd38910736c
@@ -0,0 +1,63 @@
+// SPDX-License-Identifier: GPL-2.0
+
+//! Advanced Configuration and Power Interface abstractions.
+
+use crate::{bindings, device_id::RawDeviceId, prelude::*};
+
+/// IdTable type for ACPI drivers.
+pub type IdTable<T> = &'static dyn kernel::device_id::IdTable<DeviceId, T>;
+
+/// An ACPI device id.
+#[repr(transparent)]
+#[derive(Clone, Copy)]
+pub struct DeviceId(bindings::acpi_device_id);
+
+// SAFETY:
+// * `DeviceId` is a `#[repr(transparent)` wrapper of `struct acpi_device_id` and does not add
+//   additional invariants, so it's safe to transmute to `RawType`.
+// * `DRIVER_DATA_OFFSET` is the offset to the `data` field.
+unsafe impl RawDeviceId for DeviceId {
+    type RawType = bindings::acpi_device_id;
+
+    const DRIVER_DATA_OFFSET: usize = core::mem::offset_of!(bindings::acpi_device_id, driver_data);
+
+    fn index(&self) -> usize {
+        self.0.driver_data as _
+    }
+}
+

rust/kernel/acpi.rs line 38

Quoted 10 lines
rust/kernel/acpi.rs new file +10 -0
+impl DeviceId {
+    const ACPI_ID_LEN: usize = 16;
+
+    /// Create a new device id from an ACPI 'id' string.
+    pub const fn new(id: &'static CStr) -> Self {
+        assert!(id.len() <= Self::ACPI_ID_LEN, "ID exceeds 16 bytes");
+        let src = id.as_bytes_with_nul();
+        // Replace with `bindings::acpi_device_id::default()` once stabilized for `const`.
+        // SAFETY: FFI type is valid to be zero-initialized.
+        let mut acpi: bindings::acpi_device_id = unsafe { core::mem::zeroed() };

The patch did not land yet, but for this there's the following improvement upcoming:

https://lore.kernel.org/rust-for-linux/[email protected]/

rust/kernel/acpi.rs line 39

Quoted 1 line
rust/kernel/acpi.rs new file +1 -0
+        // TODO: Use `clone_from_slice` once the corresponding types do match.

clone_from_slice() can't be used from const context, this comment is wrong in of.rs (where you probably copied it from).

rust/kernel/acpi.rs line 42

Quoted 3 lines
rust/kernel/acpi.rs new file +3 -0
+        let mut i = 0;
+        while i < src.len() {
+            acpi.id[i] = src[i] as _;

You can drop this cast, it shouldn't be needed anymore.

Quoted 26 lines
rust/kernel/acpi.rs new file +21 -0
+            i += 1;
+        }
+
+        Self(acpi)
+    }
+}
+
+/// Create an ACPI `IdTable` with an "alias" for modpost.
+#[macro_export]
+macro_rules! acpi_device_table {
+    ($table_name:ident, $module_table_name:ident, $id_info_type: ty, $table_data: expr) => {
+        const $table_name: $crate::device_id::IdArray<
+            $crate::acpi::DeviceId,
+            $id_info_type,
+            { $table_data.len() },
+        > = $crate::device_id::IdArray::new($table_data);
+
+        $crate::module_device_table!("acpi", $module_table_name, $table_name);
+    };
+}
+
rust/kernel/lib.rs +0 -0
index 15c5f72976cd..05f1d3870bf7 100644

rust/kernel/lib.rs line 81

Quoted 4 lines
rust/kernel/lib.rs +1 -0
@@ -78,6 +78,7 @@
 #[cfg(CONFIG_NET)]
 pub mod net;
 pub mod of;
+pub mod acpi;

As mentioned by Greg, please keep this ordered.

#3 lore raw
Quoted 10 lines · Danilo Krummrich
Quoted 6 lines · Igor Korotin

acpi::DeviceId is an abstraction around struct acpi_device_id.

This is used by subsequent patches, in particular the i2c driver abstractions, to create ACPI device ID tables.

As Greg mentioned it would be nice to see the subsequent patches.

Ok. There's a little misunderstanding from my side. I'm in the progress of implementation
of I2C driver abstractions. I2C drivers can use either "of" or "acpi". The idea was to push this
change first, because:
- It's quite standalone one.
- I'm not sure how much time it will take me to finalize I2C drivers
abstractions.

If it is not appropriate way of commits, I'll then keep it until all is done.

Noted all the remarks. Will sort them in the next drop.

Thanks for the review.

Thanks
Igor

#4 lore raw
Quoted 12 lines · Igor Korotin
Quoted 10 lines · Danilo Krummrich
Quoted 6 lines · Igor Korotin

acpi::DeviceId is an abstraction around struct acpi_device_id.

This is used by subsequent patches, in particular the i2c driver abstractions, to create ACPI device ID tables.

As Greg mentioned it would be nice to see the subsequent patches.

Quoted 8 lines

Ok. There's a little misunderstanding from my side. I'm in the progress of implementation
of I2C driver abstractions. I2C drivers can use either "of" or "acpi". The idea was to push this
change first, because:
- It's quite standalone one.
- I'm not sure how much time it will take me to finalize I2C drivers
abstractions.

If you don't need it now, then there's no rush to get it merged now :)

Quoted 1 line

If it is not appropriate way of commits, I'll then keep it until all is done.

We would like to see it be used first, to ensure that the code is actually correct.

thanks,

greg k-h

#5 lore raw
Quoted 22 lines · Greg KH
Quoted 12 lines · Igor Korotin
Quoted 10 lines · Danilo Krummrich
Quoted 6 lines · Igor Korotin

acpi::DeviceId is an abstraction around struct acpi_device_id.

This is used by subsequent patches, in particular the i2c driver abstractions, to create ACPI device ID tables.

As Greg mentioned it would be nice to see the subsequent patches.

Quoted 8 lines

Ok. There's a little misunderstanding from my side. I'm in the progress of implementation
of I2C driver abstractions. I2C drivers can use either "of" or "acpi". The idea was to push this
change first, because:
- It's quite standalone one.
- I'm not sure how much time it will take me to finalize I2C drivers
abstractions.

Quoted 6 lines

If you don't need it now, then there's no rush to get it merged now :)

Quoted 1 line

If it is not appropriate way of commits, I'll then keep it until all is done.

We would like to see it be used first, to ensure that the code is actually correct.

Alternatively, if you want to upstream this dependency already you can send the following patches:

- this acpi::DeviceId abstraction
- the glue code for the generic adapter trait in rust/kernel/driver.rs
- use this glue code in the platform abstraction
- add acpi support to the platform sample driver

This way we can already validate that the code works correctly. All this is required anyways if the I2C device you write a driver for is on the platform bus.

#6 lore raw
Quoted 11 lines

Alternatively, if you want to upstream this dependency already you can send the following patches:

- this acpi::DeviceId abstraction
- the glue code for the generic adapter trait in rust/kernel/driver.rs
- use this glue code in the platform abstraction
- add acpi support to the platform sample driver

This way we can already validate that the code works correctly. All this is required anyways if the I2C device you write a driver for is on the platform bus.

A few questions if I may:
1. I committed to 4 different files: acpi.rs, driver.rs, platform.rs, platform rust sample driver.
Should I commit all of this as one commit or split each part to a separate commit and send it as a patch sequence?
2. From author's point of view, as Danilo noticed, acpi table abstraction code is in general just copy-paste from of table abstraction code. How should I explicitly mark that fact?

Thanks
Igor

#7 lore raw
Quoted 12 lines · Igor Korotin
Quoted 11 lines

Alternatively, if you want to upstream this dependency already you can send the following patches:

- this acpi::DeviceId abstraction
- the glue code for the generic adapter trait in rust/kernel/driver.rs
- use this glue code in the platform abstraction
- add acpi support to the platform sample driver

This way we can already validate that the code works correctly. All this is required anyways if the I2C device you write a driver for is on the platform bus.

Quoted 5 lines

A few questions if I may:
1. I committed to 4 different files: acpi.rs, driver.rs, platform.rs, platform rust sample driver.
Should I commit all of this as one commit or split each part to a separate commit and send it as a patch sequence?

Every entry of my list above should be a separate commit. It might happen that writing the glue code for the generic adapter trait in rust/kernel/driver.rs breaks the build in the platform abstraction, then you have to fix it up in the same commit, i.e. we never break the build. Please also see [1].

Quoted 3 lines

2. From author's point of view, as Danilo noticed, acpi table abstraction code is in general just copy-paste from of table abstraction code. How should I explicitly mark that fact?

You don't need to do anything specific here. You authored the commit, even though it's based on existing code.

If you want you can add a note in the commit message that your case is based on the OF table abstraction. But AFAIC, you don't have to.

[1] https://docs.kernel.org/process/submitting-patches.html#separate-your-changes