* [PATCH v6] rust: ACPI: fix missing match data for PRP0001
@ 2026-04-27 17:55 Markus Probst
2026-04-28 17:55 ` Rafael J. Wysocki
2026-04-28 18:54 ` Danilo Krummrich
0 siblings, 2 replies; 3+ messages in thread
From: Markus Probst @ 2026-04-27 17:55 UTC (permalink / raw)
To: Rafael J. Wysocki, Len Brown, Miguel Ojeda, Boqun Feng, Gary Guo,
Björn Roy Baron, Benno Lossin, Andreas Hindborg, Alice Ryhl,
Trevor Gross, Danilo Krummrich, Greg Kroah-Hartman, Saket Dumbre
Cc: linux-kernel, linux-acpi, rust-for-linux, driver-core,
acpica-devel, Markus Probst
Export `acpi_of_match_device` function and use it to match the of device
table against ACPI PRP0001 in Rust.
This fixes id_info being None on ACPI PRP0001 devices.
Using `device_get_match_data` is not possible, because Rust stores an
index in the of device id instead of a data pointer. This was done this
way to provide a convenient and obvious API for drivers, which can be
evaluated in const context without the use of any unstable language
features.
Fixes: 7a718a1f26d1 ("rust: driver: implement `Adapter`")
Signed-off-by: Markus Probst <markus.probst@posteo.de>
---
Changes in v6:
- rebased onto v7.1-rc1
- Link to v5: https://patch.msgid.link/20260420-rust_acpi_prp0001-v5-1-f77869b18b9f@posteo.de
Changes in v5:
- provide a stub instead of canditionalize the caller
- Link to v4: https://lore.kernel.org/r/20260408-rust_acpi_prp0001-v4-1-87b78ad79239@posteo.de
Changes in v4:
- fix compilation on !CONFIG_ACPI
- fix typo
- add doc to `acpi_of_match_device` and remove from header file include
- Link to v3: https://lore.kernel.org/r/20260407-rust_acpi_prp0001-v3-1-c5b24590c273@posteo.de
Changes in v3:
- use acpi/acpi_bus.h instead of drivers/acpi/internal.h to declare
`acpi_of_match_device`
- Link to v2: https://lore.kernel.org/r/20260405-rust_acpi_prp0001-v2-1-4f8b1cf075d3@posteo.de
Changes in v2:
- keep unsafe block as short as possible
- don't export acpi_of_match_device
- add implementation decisions to commit msg
- Link to v1: https://lore.kernel.org/r/20260401-rust_acpi_prp0001-v1-1-f6a4d2ef9244@posteo.de
---
MAINTAINERS | 1 +
drivers/acpi/bus.c | 6 ++--
include/acpi/acpi_bus.h | 11 ++++++++
rust/helpers/acpi.c | 16 +++++++++++
rust/helpers/helpers.c | 1 +
rust/kernel/driver.rs | 74 ++++++++++++++++++++++++++++++++++++++++---------
6 files changed, 93 insertions(+), 16 deletions(-)
diff --git a/MAINTAINERS b/MAINTAINERS
index 2fb1c75afd16..4367a303e90e 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -291,6 +291,7 @@ F: include/linux/acpi.h
F: include/linux/fwnode.h
F: include/linux/fw_table.h
F: lib/fw_table.c
+F: rust/helpers/acpi.c
F: rust/kernel/acpi.rs
F: tools/power/acpi/
diff --git a/drivers/acpi/bus.c b/drivers/acpi/bus.c
index 2ec095e2009e..554dbddb1c8c 100644
--- a/drivers/acpi/bus.c
+++ b/drivers/acpi/bus.c
@@ -831,9 +831,9 @@ const struct acpi_device *acpi_companion_match(const struct device *dev)
* identifiers and a _DSD object with the "compatible" property, use that
* property to match against the given list of identifiers.
*/
-static bool acpi_of_match_device(const struct acpi_device *adev,
- const struct of_device_id *of_match_table,
- const struct of_device_id **of_id)
+bool acpi_of_match_device(const struct acpi_device *adev,
+ const struct of_device_id *of_match_table,
+ const struct of_device_id **of_id)
{
const union acpi_object *of_compatible, *obj;
int i, nval;
diff --git a/include/acpi/acpi_bus.h b/include/acpi/acpi_bus.h
index b701b5f972cb..c85f7d8f1632 100644
--- a/include/acpi/acpi_bus.h
+++ b/include/acpi/acpi_bus.h
@@ -187,6 +187,10 @@ struct acpi_driver {
* -----------
*/
+bool acpi_of_match_device(const struct acpi_device *adev,
+ const struct of_device_id *of_match_table,
+ const struct of_device_id **of_id);
+
/* Status (_STA) */
struct acpi_device_status {
@@ -992,6 +996,13 @@ int acpi_scan_add_dep(acpi_handle handle, struct acpi_handle_list *dep_devices);
u32 arch_acpi_add_auto_dep(acpi_handle handle);
#else /* CONFIG_ACPI */
+static inline bool acpi_of_match_device(const struct acpi_device *adev,
+ const struct of_device_id *of_match_table,
+ const struct of_device_id **of_id)
+{
+ return false;
+}
+
static inline int register_acpi_bus_type(void *bus) { return 0; }
static inline int unregister_acpi_bus_type(void *bus) { return 0; }
diff --git a/rust/helpers/acpi.c b/rust/helpers/acpi.c
new file mode 100644
index 000000000000..e75c9807bbad
--- /dev/null
+++ b/rust/helpers/acpi.c
@@ -0,0 +1,16 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include <linux/acpi.h>
+#include <acpi/acpi_bus.h>
+
+__rust_helper bool rust_helper_acpi_of_match_device(const struct acpi_device *adev,
+ const struct of_device_id *of_match_table,
+ const struct of_device_id **of_id)
+{
+ return acpi_of_match_device(adev, of_match_table, of_id);
+}
+
+__rust_helper struct acpi_device *rust_helper_to_acpi_device_node(struct fwnode_handle *fwnode)
+{
+ return to_acpi_device_node(fwnode);
+}
diff --git a/rust/helpers/helpers.c b/rust/helpers/helpers.c
index 625921e27dfb..38b34518eff1 100644
--- a/rust/helpers/helpers.c
+++ b/rust/helpers/helpers.c
@@ -38,6 +38,7 @@
#define __rust_helper __always_inline
#endif
+#include "acpi.c"
#include "atomic.c"
#include "atomic_ext.c"
#include "auxiliary.c"
diff --git a/rust/kernel/driver.rs b/rust/kernel/driver.rs
index 36de8098754d..93e5dd6ae371 100644
--- a/rust/kernel/driver.rs
+++ b/rust/kernel/driver.rs
@@ -278,6 +278,26 @@ fn init(
}
}
+// Calling the FFI function directly from the `Adapter` impl may result in it being called
+// directly from driver modules. This happens since the Rust compiler will use monomorphisation, so
+// it might happen that functions are instantiated within the calling driver module. For now, work
+// around this with `#[inline(never)]` helpers.
+//
+// TODO: Remove once a more generic solution has been implemented. For instance, we may be able to
+// leverage `bindgen` to take care of this depending on whether a symbol is (already) exported.
+#[inline(never)]
+#[allow(clippy::missing_safety_doc)]
+#[allow(dead_code)]
+#[must_use]
+unsafe fn acpi_of_match_device(
+ adev: *const bindings::acpi_device,
+ of_match_table: *const bindings::of_device_id,
+ of_id: *mut *const bindings::of_device_id,
+) -> bool {
+ // SAFETY: Safety requirements are the same as `bindings::acpi_of_match_device`.
+ unsafe { bindings::acpi_of_match_device(adev, of_match_table, of_id) }
+}
+
/// The bus independent adapter to match a drivers and a devices.
///
/// This trait should be implemented by the bus specific adapter, which represents the connection
@@ -329,35 +349,63 @@ fn acpi_id_info(dev: &device::Device) -> Option<&'static Self::IdInfo> {
///
/// If this returns `None`, it means there is no match with an entry in the [`of::IdTable`].
fn of_id_info(dev: &device::Device) -> Option<&'static Self::IdInfo> {
- #[cfg(not(CONFIG_OF))]
+ let table = Self::of_id_table()?;
+
+ #[cfg(not(any(CONFIG_OF, CONFIG_ACPI)))]
{
- let _ = dev;
- None
+ let _ = (dev, table);
}
#[cfg(CONFIG_OF)]
{
- let table = Self::of_id_table()?;
-
// SAFETY:
// - `table` has static lifetime, hence it's valid for read,
// - `dev` is guaranteed to be valid while it's alive, and so is `dev.as_raw()`.
let raw_id = unsafe { bindings::of_match_device(table.as_ptr(), dev.as_raw()) };
- if raw_id.is_null() {
- None
- } else {
+ if !raw_id.is_null() {
// SAFETY: `DeviceId` is a `#[repr(transparent)]` wrapper of `struct of_device_id`
// and does not add additional invariants, so it's safe to transmute.
let id = unsafe { &*raw_id.cast::<of::DeviceId>() };
- Some(
- table.info(<of::DeviceId as crate::device_id::RawDeviceIdIndex>::index(
- id,
- )),
- )
+ return Some(table.info(
+ <of::DeviceId as crate::device_id::RawDeviceIdIndex>::index(id),
+ ));
}
}
+
+ #[cfg(CONFIG_ACPI)]
+ {
+ use core::ptr;
+ use device::property::FwNode;
+
+ let mut raw_id = ptr::null();
+
+ let fwnode = dev.fwnode().map_or(ptr::null_mut(), FwNode::as_raw);
+
+ // SAFETY: `fwnode` is a pointer to a valid `fwnode_handle`. A null pointer will be
+ // passed through the function.
+ let adev = unsafe { bindings::to_acpi_device_node(fwnode) };
+
+ // SAFETY:
+ // - `adev` is a valid pointer to `acpi_device` or is null. It is guaranteed to be
+ // valid as long as `dev` is alive.
+ // - `table` has static lifetime, hence it's valid for read.
+ if unsafe { acpi_of_match_device(adev, table.as_ptr(), &raw mut raw_id) } {
+ // SAFETY:
+ // - the function returns true, therefore `raw_id` has been set to a pointer to a
+ // valid `of_device_id`.
+ // - `DeviceId` is a `#[repr(transparent)]` wrapper of `struct of_device_id`
+ // and does not add additional invariants, so it's safe to transmute.
+ let id = unsafe { &*raw_id.cast::<of::DeviceId>() };
+
+ return Some(table.info(
+ <of::DeviceId as crate::device_id::RawDeviceIdIndex>::index(id),
+ ));
+ }
+ }
+
+ None
}
/// Returns the driver's private data from the matching entry of any of the ID tables, if any.
---
base-commit: 3131ff5a117498bb4b9db3a238bb311cbf8383ce
change-id: 20260401-rust_acpi_prp0001-a2971543b555
^ permalink raw reply related [flat|nested] 3+ messages in thread* Re: [PATCH v6] rust: ACPI: fix missing match data for PRP0001
2026-04-27 17:55 [PATCH v6] rust: ACPI: fix missing match data for PRP0001 Markus Probst
@ 2026-04-28 17:55 ` Rafael J. Wysocki
2026-04-28 18:54 ` Danilo Krummrich
1 sibling, 0 replies; 3+ messages in thread
From: Rafael J. Wysocki @ 2026-04-28 17:55 UTC (permalink / raw)
To: Markus Probst
Cc: Rafael J. Wysocki, Len Brown, Miguel Ojeda, Boqun Feng, Gary Guo,
Björn Roy Baron, Benno Lossin, Andreas Hindborg, Alice Ryhl,
Trevor Gross, Danilo Krummrich, Greg Kroah-Hartman, Saket Dumbre,
linux-kernel, linux-acpi, rust-for-linux, driver-core,
acpica-devel
On Mon, Apr 27, 2026 at 7:56 PM Markus Probst <markus.probst@posteo.de> wrote:
>
> Export `acpi_of_match_device` function and use it to match the of device
> table against ACPI PRP0001 in Rust.
>
> This fixes id_info being None on ACPI PRP0001 devices.
>
> Using `device_get_match_data` is not possible, because Rust stores an
> index in the of device id instead of a data pointer. This was done this
> way to provide a convenient and obvious API for drivers, which can be
> evaluated in const context without the use of any unstable language
> features.
>
> Fixes: 7a718a1f26d1 ("rust: driver: implement `Adapter`")
> Signed-off-by: Markus Probst <markus.probst@posteo.de>
Acked-by: Rafael J. Wysocki (Intel) <rafael@kernel.org> # ACPI
(and I think I've already ACKed it once).
> ---
> Changes in v6:
> - rebased onto v7.1-rc1
> - Link to v5: https://patch.msgid.link/20260420-rust_acpi_prp0001-v5-1-f77869b18b9f@posteo.de
>
> Changes in v5:
> - provide a stub instead of canditionalize the caller
> - Link to v4: https://lore.kernel.org/r/20260408-rust_acpi_prp0001-v4-1-87b78ad79239@posteo.de
>
> Changes in v4:
> - fix compilation on !CONFIG_ACPI
> - fix typo
> - add doc to `acpi_of_match_device` and remove from header file include
> - Link to v3: https://lore.kernel.org/r/20260407-rust_acpi_prp0001-v3-1-c5b24590c273@posteo.de
>
> Changes in v3:
> - use acpi/acpi_bus.h instead of drivers/acpi/internal.h to declare
> `acpi_of_match_device`
> - Link to v2: https://lore.kernel.org/r/20260405-rust_acpi_prp0001-v2-1-4f8b1cf075d3@posteo.de
>
> Changes in v2:
> - keep unsafe block as short as possible
> - don't export acpi_of_match_device
> - add implementation decisions to commit msg
> - Link to v1: https://lore.kernel.org/r/20260401-rust_acpi_prp0001-v1-1-f6a4d2ef9244@posteo.de
> ---
> MAINTAINERS | 1 +
> drivers/acpi/bus.c | 6 ++--
> include/acpi/acpi_bus.h | 11 ++++++++
> rust/helpers/acpi.c | 16 +++++++++++
> rust/helpers/helpers.c | 1 +
> rust/kernel/driver.rs | 74 ++++++++++++++++++++++++++++++++++++++++---------
> 6 files changed, 93 insertions(+), 16 deletions(-)
>
> diff --git a/MAINTAINERS b/MAINTAINERS
> index 2fb1c75afd16..4367a303e90e 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -291,6 +291,7 @@ F: include/linux/acpi.h
> F: include/linux/fwnode.h
> F: include/linux/fw_table.h
> F: lib/fw_table.c
> +F: rust/helpers/acpi.c
> F: rust/kernel/acpi.rs
> F: tools/power/acpi/
>
> diff --git a/drivers/acpi/bus.c b/drivers/acpi/bus.c
> index 2ec095e2009e..554dbddb1c8c 100644
> --- a/drivers/acpi/bus.c
> +++ b/drivers/acpi/bus.c
> @@ -831,9 +831,9 @@ const struct acpi_device *acpi_companion_match(const struct device *dev)
> * identifiers and a _DSD object with the "compatible" property, use that
> * property to match against the given list of identifiers.
> */
> -static bool acpi_of_match_device(const struct acpi_device *adev,
> - const struct of_device_id *of_match_table,
> - const struct of_device_id **of_id)
> +bool acpi_of_match_device(const struct acpi_device *adev,
> + const struct of_device_id *of_match_table,
> + const struct of_device_id **of_id)
> {
> const union acpi_object *of_compatible, *obj;
> int i, nval;
> diff --git a/include/acpi/acpi_bus.h b/include/acpi/acpi_bus.h
> index b701b5f972cb..c85f7d8f1632 100644
> --- a/include/acpi/acpi_bus.h
> +++ b/include/acpi/acpi_bus.h
> @@ -187,6 +187,10 @@ struct acpi_driver {
> * -----------
> */
>
> +bool acpi_of_match_device(const struct acpi_device *adev,
> + const struct of_device_id *of_match_table,
> + const struct of_device_id **of_id);
> +
> /* Status (_STA) */
>
> struct acpi_device_status {
> @@ -992,6 +996,13 @@ int acpi_scan_add_dep(acpi_handle handle, struct acpi_handle_list *dep_devices);
> u32 arch_acpi_add_auto_dep(acpi_handle handle);
> #else /* CONFIG_ACPI */
>
> +static inline bool acpi_of_match_device(const struct acpi_device *adev,
> + const struct of_device_id *of_match_table,
> + const struct of_device_id **of_id)
> +{
> + return false;
> +}
> +
> static inline int register_acpi_bus_type(void *bus) { return 0; }
> static inline int unregister_acpi_bus_type(void *bus) { return 0; }
>
> diff --git a/rust/helpers/acpi.c b/rust/helpers/acpi.c
> new file mode 100644
> index 000000000000..e75c9807bbad
> --- /dev/null
> +++ b/rust/helpers/acpi.c
> @@ -0,0 +1,16 @@
> +// SPDX-License-Identifier: GPL-2.0
> +
> +#include <linux/acpi.h>
> +#include <acpi/acpi_bus.h>
> +
> +__rust_helper bool rust_helper_acpi_of_match_device(const struct acpi_device *adev,
> + const struct of_device_id *of_match_table,
> + const struct of_device_id **of_id)
> +{
> + return acpi_of_match_device(adev, of_match_table, of_id);
> +}
> +
> +__rust_helper struct acpi_device *rust_helper_to_acpi_device_node(struct fwnode_handle *fwnode)
> +{
> + return to_acpi_device_node(fwnode);
> +}
> diff --git a/rust/helpers/helpers.c b/rust/helpers/helpers.c
> index 625921e27dfb..38b34518eff1 100644
> --- a/rust/helpers/helpers.c
> +++ b/rust/helpers/helpers.c
> @@ -38,6 +38,7 @@
> #define __rust_helper __always_inline
> #endif
>
> +#include "acpi.c"
> #include "atomic.c"
> #include "atomic_ext.c"
> #include "auxiliary.c"
> diff --git a/rust/kernel/driver.rs b/rust/kernel/driver.rs
> index 36de8098754d..93e5dd6ae371 100644
> --- a/rust/kernel/driver.rs
> +++ b/rust/kernel/driver.rs
> @@ -278,6 +278,26 @@ fn init(
> }
> }
>
> +// Calling the FFI function directly from the `Adapter` impl may result in it being called
> +// directly from driver modules. This happens since the Rust compiler will use monomorphisation, so
> +// it might happen that functions are instantiated within the calling driver module. For now, work
> +// around this with `#[inline(never)]` helpers.
> +//
> +// TODO: Remove once a more generic solution has been implemented. For instance, we may be able to
> +// leverage `bindgen` to take care of this depending on whether a symbol is (already) exported.
> +#[inline(never)]
> +#[allow(clippy::missing_safety_doc)]
> +#[allow(dead_code)]
> +#[must_use]
> +unsafe fn acpi_of_match_device(
> + adev: *const bindings::acpi_device,
> + of_match_table: *const bindings::of_device_id,
> + of_id: *mut *const bindings::of_device_id,
> +) -> bool {
> + // SAFETY: Safety requirements are the same as `bindings::acpi_of_match_device`.
> + unsafe { bindings::acpi_of_match_device(adev, of_match_table, of_id) }
> +}
> +
> /// The bus independent adapter to match a drivers and a devices.
> ///
> /// This trait should be implemented by the bus specific adapter, which represents the connection
> @@ -329,35 +349,63 @@ fn acpi_id_info(dev: &device::Device) -> Option<&'static Self::IdInfo> {
> ///
> /// If this returns `None`, it means there is no match with an entry in the [`of::IdTable`].
> fn of_id_info(dev: &device::Device) -> Option<&'static Self::IdInfo> {
> - #[cfg(not(CONFIG_OF))]
> + let table = Self::of_id_table()?;
> +
> + #[cfg(not(any(CONFIG_OF, CONFIG_ACPI)))]
> {
> - let _ = dev;
> - None
> + let _ = (dev, table);
> }
>
> #[cfg(CONFIG_OF)]
> {
> - let table = Self::of_id_table()?;
> -
> // SAFETY:
> // - `table` has static lifetime, hence it's valid for read,
> // - `dev` is guaranteed to be valid while it's alive, and so is `dev.as_raw()`.
> let raw_id = unsafe { bindings::of_match_device(table.as_ptr(), dev.as_raw()) };
>
> - if raw_id.is_null() {
> - None
> - } else {
> + if !raw_id.is_null() {
> // SAFETY: `DeviceId` is a `#[repr(transparent)]` wrapper of `struct of_device_id`
> // and does not add additional invariants, so it's safe to transmute.
> let id = unsafe { &*raw_id.cast::<of::DeviceId>() };
>
> - Some(
> - table.info(<of::DeviceId as crate::device_id::RawDeviceIdIndex>::index(
> - id,
> - )),
> - )
> + return Some(table.info(
> + <of::DeviceId as crate::device_id::RawDeviceIdIndex>::index(id),
> + ));
> }
> }
> +
> + #[cfg(CONFIG_ACPI)]
> + {
> + use core::ptr;
> + use device::property::FwNode;
> +
> + let mut raw_id = ptr::null();
> +
> + let fwnode = dev.fwnode().map_or(ptr::null_mut(), FwNode::as_raw);
> +
> + // SAFETY: `fwnode` is a pointer to a valid `fwnode_handle`. A null pointer will be
> + // passed through the function.
> + let adev = unsafe { bindings::to_acpi_device_node(fwnode) };
> +
> + // SAFETY:
> + // - `adev` is a valid pointer to `acpi_device` or is null. It is guaranteed to be
> + // valid as long as `dev` is alive.
> + // - `table` has static lifetime, hence it's valid for read.
> + if unsafe { acpi_of_match_device(adev, table.as_ptr(), &raw mut raw_id) } {
> + // SAFETY:
> + // - the function returns true, therefore `raw_id` has been set to a pointer to a
> + // valid `of_device_id`.
> + // - `DeviceId` is a `#[repr(transparent)]` wrapper of `struct of_device_id`
> + // and does not add additional invariants, so it's safe to transmute.
> + let id = unsafe { &*raw_id.cast::<of::DeviceId>() };
> +
> + return Some(table.info(
> + <of::DeviceId as crate::device_id::RawDeviceIdIndex>::index(id),
> + ));
> + }
> + }
> +
> + None
> }
>
> /// Returns the driver's private data from the matching entry of any of the ID tables, if any.
>
> ---
> base-commit: 3131ff5a117498bb4b9db3a238bb311cbf8383ce
> change-id: 20260401-rust_acpi_prp0001-a2971543b555
>
>
^ permalink raw reply [flat|nested] 3+ messages in thread* Re: [PATCH v6] rust: ACPI: fix missing match data for PRP0001
2026-04-27 17:55 [PATCH v6] rust: ACPI: fix missing match data for PRP0001 Markus Probst
2026-04-28 17:55 ` Rafael J. Wysocki
@ 2026-04-28 18:54 ` Danilo Krummrich
1 sibling, 0 replies; 3+ messages in thread
From: Danilo Krummrich @ 2026-04-28 18:54 UTC (permalink / raw)
To: Markus Probst
Cc: Rafael J . Wysocki, Len Brown, Miguel Ojeda, Boqun Feng, Gary Guo,
Björn Roy Baron, Benno Lossin, Andreas Hindborg, Alice Ryhl,
Trevor Gross, Danilo Krummrich, Greg Kroah-Hartman, Saket Dumbre,
linux-kernel, linux-acpi, rust-for-linux, driver-core,
acpica-devel
On Mon, 27 Apr 2026 17:55:57 +0000, Markus Probst wrote:
> [PATCH v6] rust: ACPI: fix missing match data for PRP0001
Applied, thanks!
Branch: driver-core-testing
Tree: git://git.kernel.org/pub/scm/linux/kernel/git/driver-core/driver-core.git
[1/1] rust: ACPI: fix missing match data for PRP0001
commit: 2690d071584e
The patch will appear in the next linux-next integration (typically within 24
hours on weekdays).
The patch is in the driver-core-testing branch and will be promoted to
driver-core-next after validation.
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2026-04-28 18:55 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-04-27 17:55 [PATCH v6] rust: ACPI: fix missing match data for PRP0001 Markus Probst
2026-04-28 17:55 ` Rafael J. Wysocki
2026-04-28 18:54 ` Danilo Krummrich
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox