* [PATCH v7 1/7] rust: drm: gem: Add raw_dma_resv() function
2026-02-06 22:34 [PATCH v7 0/7] Rust bindings for gem shmem + iosys_map Lyude Paul
@ 2026-02-06 22:34 ` Lyude Paul
2026-02-09 12:40 ` Alice Ryhl
2026-02-06 22:34 ` [PATCH v7 2/7] rust: helpers: Add bindings/wrappers for dma_resv_lock Lyude Paul
` (5 subsequent siblings)
6 siblings, 1 reply; 16+ messages in thread
From: Lyude Paul @ 2026-02-06 22:34 UTC (permalink / raw)
To: linux-kernel, dri-devel, rust-for-linux, Danilo Krummrich
Cc: nouveau, Daniel Almeida, Gary Guo, Benno Lossin,
Alexandre Courbot, Janne Grunau
For retrieving a pointer to the struct dma_resv for a given GEM object. We
also introduce it in a new trait, BaseObjectPrivate, which we automatically
implement for all gem objects and don't expose to users outside of the
crate.
Signed-off-by: Lyude Paul <lyude@redhat.com>
Reviewed-by: Janne Grunau <j@jananu.net>
---
rust/kernel/drm/gem/mod.rs | 12 ++++++++++++
1 file changed, 12 insertions(+)
diff --git a/rust/kernel/drm/gem/mod.rs b/rust/kernel/drm/gem/mod.rs
index e1cc3af8ade9b..82b3151e5ae3b 100644
--- a/rust/kernel/drm/gem/mod.rs
+++ b/rust/kernel/drm/gem/mod.rs
@@ -197,6 +197,18 @@ fn create_mmap_offset(&self) -> Result<u64> {
impl<T: IntoGEMObject> BaseObject for T {}
+/// Crate-private base operations shared by all GEM object classes.
+#[expect(unused)]
+pub(crate) trait BaseObjectPrivate: IntoGEMObject {
+ /// Return a pointer to this object's dma_resv.
+ fn raw_dma_resv(&self) -> *mut bindings::dma_resv {
+ // SAFETY: `as_gem_obj()` always returns a valid pointer to the base DRM gem object
+ unsafe { (*self.as_raw()).resv }
+ }
+}
+
+impl<T: IntoGEMObject> BaseObjectPrivate for T {}
+
/// A base GEM object.
///
/// # Invariants
--
2.53.0
^ permalink raw reply related [flat|nested] 16+ messages in thread* Re: [PATCH v7 1/7] rust: drm: gem: Add raw_dma_resv() function
2026-02-06 22:34 ` [PATCH v7 1/7] rust: drm: gem: Add raw_dma_resv() function Lyude Paul
@ 2026-02-09 12:40 ` Alice Ryhl
2026-02-09 12:47 ` Janne Grunau
0 siblings, 1 reply; 16+ messages in thread
From: Alice Ryhl @ 2026-02-09 12:40 UTC (permalink / raw)
To: Lyude Paul
Cc: linux-kernel, dri-devel, rust-for-linux, Danilo Krummrich,
nouveau, Daniel Almeida, Gary Guo, Benno Lossin,
Alexandre Courbot, Janne Grunau
On Fri, Feb 06, 2026 at 05:34:25PM -0500, Lyude Paul wrote:
> Reviewed-by: Janne Grunau <j@jananu.net>
I think this email is misspelled?
^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH v7 1/7] rust: drm: gem: Add raw_dma_resv() function
2026-02-09 12:40 ` Alice Ryhl
@ 2026-02-09 12:47 ` Janne Grunau
0 siblings, 0 replies; 16+ messages in thread
From: Janne Grunau @ 2026-02-09 12:47 UTC (permalink / raw)
To: Alice Ryhl, Lyude Paul
Cc: LKML, dri-devel, rust-for-linux, Danilo Krummrich, nouveau,
Daniel Almeida, Gary Guo, Benno Lossin, Alexandre Courbot
On Mon, Feb 9, 2026, at 13:40, Alice Ryhl wrote:
> On Fri, Feb 06, 2026 at 05:34:25PM -0500, Lyude Paul wrote:
>> Reviewed-by: Janne Grunau <j@jananu.net>
>
> I think this email is misspelled?
It is, likely my fault. I noticed a few of those particular misspellings in the last few weeks.
Janne
^ permalink raw reply [flat|nested] 16+ messages in thread
* [PATCH v7 2/7] rust: helpers: Add bindings/wrappers for dma_resv_lock
2026-02-06 22:34 [PATCH v7 0/7] Rust bindings for gem shmem + iosys_map Lyude Paul
2026-02-06 22:34 ` [PATCH v7 1/7] rust: drm: gem: Add raw_dma_resv() function Lyude Paul
@ 2026-02-06 22:34 ` Lyude Paul
2026-02-06 22:34 ` [PATCH v7 3/7] rust: gem: Introduce DriverObject::Args Lyude Paul
` (4 subsequent siblings)
6 siblings, 0 replies; 16+ messages in thread
From: Lyude Paul @ 2026-02-06 22:34 UTC (permalink / raw)
To: linux-kernel, dri-devel, rust-for-linux, Danilo Krummrich
Cc: nouveau, Daniel Almeida, Gary Guo, Benno Lossin,
Alexandre Courbot, Janne Grunau
From: Asahi Lina <lina@asahilina.net>
This is just for basic usage in the DRM shmem abstractions for implied
locking, not intended as a full DMA Reservation abstraction yet.
Signed-off-by: Asahi Lina <lina@asahilina.net>
Signed-off-by: Daniel Almeida <daniel.almeida@collabora.com>
Reviewed-by: Alice Ryhl <aliceryhl@google.com>
Reviewed-by: Janne Grunau <j@jananu.net>
Signed-off-by: Lyude Paul <lyude@redhat.com>
---
rust/bindings/bindings_helper.h | 1 +
rust/helpers/dma-resv.c | 13 +++++++++++++
rust/helpers/helpers.c | 1 +
3 files changed, 15 insertions(+)
create mode 100644 rust/helpers/dma-resv.c
diff --git a/rust/bindings/bindings_helper.h b/rust/bindings/bindings_helper.h
index 9fdf76ca630e0..ecf31681df806 100644
--- a/rust/bindings/bindings_helper.h
+++ b/rust/bindings/bindings_helper.h
@@ -48,6 +48,7 @@
#include <linux/cpumask.h>
#include <linux/cred.h>
#include <linux/debugfs.h>
+#include <linux/dma-resv.h>
#include <linux/device/faux.h>
#include <linux/dma-direction.h>
#include <linux/dma-mapping.h>
diff --git a/rust/helpers/dma-resv.c b/rust/helpers/dma-resv.c
new file mode 100644
index 0000000000000..05501cb814513
--- /dev/null
+++ b/rust/helpers/dma-resv.c
@@ -0,0 +1,13 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include <linux/dma-resv.h>
+
+int rust_helper_dma_resv_lock(struct dma_resv *obj, struct ww_acquire_ctx *ctx)
+{
+ return dma_resv_lock(obj, ctx);
+}
+
+void rust_helper_dma_resv_unlock(struct dma_resv *obj)
+{
+ dma_resv_unlock(obj);
+}
diff --git a/rust/helpers/helpers.c b/rust/helpers/helpers.c
index 79c72762ad9c4..1d3333cc0d2a8 100644
--- a/rust/helpers/helpers.c
+++ b/rust/helpers/helpers.c
@@ -25,6 +25,7 @@
#include "cred.c"
#include "device.c"
#include "dma.c"
+#include "dma-resv.c"
#include "drm.c"
#include "err.c"
#include "irq.c"
--
2.53.0
^ permalink raw reply related [flat|nested] 16+ messages in thread* [PATCH v7 3/7] rust: gem: Introduce DriverObject::Args
2026-02-06 22:34 [PATCH v7 0/7] Rust bindings for gem shmem + iosys_map Lyude Paul
2026-02-06 22:34 ` [PATCH v7 1/7] rust: drm: gem: Add raw_dma_resv() function Lyude Paul
2026-02-06 22:34 ` [PATCH v7 2/7] rust: helpers: Add bindings/wrappers for dma_resv_lock Lyude Paul
@ 2026-02-06 22:34 ` Lyude Paul
2026-02-06 22:34 ` [PATCH v7 4/7] rust: drm: gem: shmem: Add DRM shmem helper abstraction Lyude Paul
` (3 subsequent siblings)
6 siblings, 0 replies; 16+ messages in thread
From: Lyude Paul @ 2026-02-06 22:34 UTC (permalink / raw)
To: linux-kernel, dri-devel, rust-for-linux, Danilo Krummrich
Cc: nouveau, Daniel Almeida, Gary Guo, Benno Lossin,
Alexandre Courbot, Janne Grunau
This is an associated type that may be used in order to specify a data-type
to pass to gem objects when construction them, allowing for drivers to more
easily initialize their private-data for gem objects.
Signed-off-by: Lyude Paul <lyude@redhat.com>
Reviewed-by: Alice Ryhl <aliceryhl@google.com>
Reviewed-by: Daniel Almeida <daniel.almeida@collabora.com>
Reviewed-by: Janne Grunau <j@jananu.net>
---
V3:
* s/BaseDriverObject/DriverObject/
V4:
* Fix leftover reference to BaseObjectDriver in rustdoc for
DriverObject::Args
V6:
* Fix build errors in Tyr
drivers/gpu/drm/nova/gem.rs | 5 +++--
drivers/gpu/drm/tyr/gem.rs | 3 ++-
rust/kernel/drm/gem/mod.rs | 13 ++++++++++---
3 files changed, 15 insertions(+), 6 deletions(-)
diff --git a/drivers/gpu/drm/nova/gem.rs b/drivers/gpu/drm/nova/gem.rs
index 2760ba4f3450b..173077eeb2def 100644
--- a/drivers/gpu/drm/nova/gem.rs
+++ b/drivers/gpu/drm/nova/gem.rs
@@ -18,8 +18,9 @@ pub(crate) struct NovaObject {}
impl gem::DriverObject for NovaObject {
type Driver = NovaDriver;
+ type Args = ();
- fn new(_dev: &NovaDevice, _size: usize) -> impl PinInit<Self, Error> {
+ fn new(_dev: &NovaDevice, _size: usize, _args: Self::Args) -> impl PinInit<Self, Error> {
try_pin_init!(NovaObject {})
}
}
@@ -33,7 +34,7 @@ pub(crate) fn new(dev: &NovaDevice, size: usize) -> Result<ARef<gem::Object<Self
return Err(EINVAL);
}
- gem::Object::new(dev, aligned_size)
+ gem::Object::new(dev, aligned_size, ())
}
/// Look up a GEM object handle for a `File` and return an `ObjectRef` for it.
diff --git a/drivers/gpu/drm/tyr/gem.rs b/drivers/gpu/drm/tyr/gem.rs
index 1273bf89dbd5d..bb5e7871efa94 100644
--- a/drivers/gpu/drm/tyr/gem.rs
+++ b/drivers/gpu/drm/tyr/gem.rs
@@ -11,8 +11,9 @@ pub(crate) struct TyrObject {}
impl gem::DriverObject for TyrObject {
type Driver = TyrDriver;
+ type Args = ();
- fn new(_dev: &TyrDevice, _size: usize) -> impl PinInit<Self, Error> {
+ fn new(_dev: &TyrDevice, _size: usize, _args: ()) -> impl PinInit<Self, Error> {
try_pin_init!(TyrObject {})
}
}
diff --git a/rust/kernel/drm/gem/mod.rs b/rust/kernel/drm/gem/mod.rs
index 82b3151e5ae3b..972d50d4342dd 100644
--- a/rust/kernel/drm/gem/mod.rs
+++ b/rust/kernel/drm/gem/mod.rs
@@ -62,8 +62,15 @@ pub trait DriverObject: Sync + Send + Sized {
/// Parent `Driver` for this object.
type Driver: drm::Driver;
+ /// The data type to use for passing arguments to [`DriverObject::new`].
+ type Args;
+
/// Create a new driver data object for a GEM object of a given size.
- fn new(dev: &drm::Device<Self::Driver>, size: usize) -> impl PinInit<Self, Error>;
+ fn new(
+ dev: &drm::Device<Self::Driver>,
+ size: usize,
+ args: Self::Args,
+ ) -> impl PinInit<Self, Error>;
/// Open a new handle to an existing object, associated with a File.
fn open(_obj: &<Self::Driver as drm::Driver>::Object, _file: &DriverFile<Self>) -> Result {
@@ -242,11 +249,11 @@ impl<T: DriverObject> Object<T> {
};
/// Create a new GEM object.
- pub fn new(dev: &drm::Device<T::Driver>, size: usize) -> Result<ARef<Self>> {
+ pub fn new(dev: &drm::Device<T::Driver>, size: usize, args: T::Args) -> Result<ARef<Self>> {
let obj: Pin<KBox<Self>> = KBox::pin_init(
try_pin_init!(Self {
obj: Opaque::new(bindings::drm_gem_object::default()),
- data <- T::new(dev, size),
+ data <- T::new(dev, size, args),
}),
GFP_KERNEL,
)?;
--
2.53.0
^ permalink raw reply related [flat|nested] 16+ messages in thread* [PATCH v7 4/7] rust: drm: gem: shmem: Add DRM shmem helper abstraction
2026-02-06 22:34 [PATCH v7 0/7] Rust bindings for gem shmem + iosys_map Lyude Paul
` (2 preceding siblings ...)
2026-02-06 22:34 ` [PATCH v7 3/7] rust: gem: Introduce DriverObject::Args Lyude Paul
@ 2026-02-06 22:34 ` Lyude Paul
2026-02-17 19:38 ` Daniel Almeida
2026-02-06 22:34 ` [PATCH v7 5/7] rust: drm: gem: Introduce shmem::SGTable Lyude Paul
` (2 subsequent siblings)
6 siblings, 1 reply; 16+ messages in thread
From: Lyude Paul @ 2026-02-06 22:34 UTC (permalink / raw)
To: linux-kernel, dri-devel, rust-for-linux, Danilo Krummrich
Cc: nouveau, Daniel Almeida, Gary Guo, Benno Lossin,
Alexandre Courbot, Janne Grunau
From: Asahi Lina <lina@asahilina.net>
The DRM shmem helper includes common code useful for drivers which
allocate GEM objects as anonymous shmem. Add a Rust abstraction for
this. Drivers can choose the raw GEM implementation or the shmem layer,
depending on their needs.
Signed-off-by: Asahi Lina <lina@asahilina.net>
Signed-off-by: Daniel Almeida <daniel.almeida@collabora.com>
Signed-off-by: Lyude Paul <lyude@redhat.com>
---
V2:
* Use the drm_gem_shmem_init() and drm_gem_shmem_release() that I extracted
so we can handle memory allocation in rust, which means we no longer have
to handle freeing rust members of the struct by hand and have a closer
implementation to the main gem object
(this also gets rid of gem_create_object)
* Get rid of GemObjectRef and UniqueGemObjectRef, we have ARef<T> at home.
* Use Device<T::Driver> in Object<T>
* Cleanup Object::<T>::new() a bit:
* Cleanup safety comment
* Use cast_mut()
* Just import container_of!(), we use it all over anyhow
* mut_shmem() -> as_shmem(), make it safe (there's no reason for being unsafe)
* Remove any *const and *muts in structs, just use NonNull
* Get rid of the previously hand-rolled sg_table bindings in shmem, use the
bindings from Abdiel's sg_table patch series
* Add a TODO at the top about DMA reservation APIs and a desire for WwMutex
* Get rid of map_wc() and replace it with a new ObjectConfig struct. While
it currently only specifies the map_wc flag, the idea here is that
settings like map_wc() and parent_resv_obj() shouldn't be exposed as
normal functions since the only place where it's safe to set them is
when we're still guaranteed unique access to the GEM object, e.g. before
returning it to the caller. Using a struct instead of individual
arguments here is mainly because we'll be adding at least one more
argument, and there's enough other gem shmem settings that trying to add
all of them as individual function arguments in the future would be a bit
messy.
* Get rid of vm_numa_fields!, Lina didn't like this macro much either and I
think that it's fine for us to just specify the #[cfg(…)] attributes by
hand since we only need to do it twice.
* Set drm_gem_object_funcs.vm_ops directly to drm_gem_shmem_vm_ops, don't
export the various shmem funcs. I'm not sure why this wasn't possible
before but it seems to work fine now.
V4:
* Switch from OpaqueObject to a normal Object<T> now that we've removed it
* Remove `dev` from Object, it's redundant as drm_gem_object already has a
device pointer we can use.
* Use &raw instead of addr_of!()
V5:
* Fix typo in shmem::Object::as_raw()
* Add type invariant around `obj` always being a valid
`drm_gem_shmem_object` for the duration of the lifetime of `Object`.
* Use Opaque::cast_from() instead of unrestricted casts
* Use IS_ENABLED() for gem shmem C helpers.
V7:
* Fix import style
* Don't forget to make Object<T> Send/Sync
rust/bindings/bindings_helper.h | 2 +
rust/helpers/drm.c | 56 ++++++-
rust/kernel/drm/gem/mod.rs | 5 +-
rust/kernel/drm/gem/shmem.rs | 249 ++++++++++++++++++++++++++++++++
4 files changed, 310 insertions(+), 2 deletions(-)
create mode 100644 rust/kernel/drm/gem/shmem.rs
diff --git a/rust/bindings/bindings_helper.h b/rust/bindings/bindings_helper.h
index ecf31681df806..c93afbd095e87 100644
--- a/rust/bindings/bindings_helper.h
+++ b/rust/bindings/bindings_helper.h
@@ -33,6 +33,7 @@
#include <drm/drm_drv.h>
#include <drm/drm_file.h>
#include <drm/drm_gem.h>
+#include <drm/drm_gem_shmem_helper.h>
#include <drm/drm_ioctl.h>
#include <kunit/test.h>
#include <linux/auxiliary_bus.h>
@@ -61,6 +62,7 @@
#include <linux/fs.h>
#include <linux/i2c.h>
#include <linux/ioport.h>
+#include <linux/iosys-map.h>
#include <linux/jiffies.h>
#include <linux/jump_label.h>
#include <linux/mdio.h>
diff --git a/rust/helpers/drm.c b/rust/helpers/drm.c
index 450b406c6f273..53ec06879db3d 100644
--- a/rust/helpers/drm.c
+++ b/rust/helpers/drm.c
@@ -1,6 +1,7 @@
// SPDX-License-Identifier: GPL-2.0
#include <drm/drm_gem.h>
+#include <drm/drm_gem_shmem_helper.h>
#include <drm/drm_vma_manager.h>
#ifdef CONFIG_DRM
@@ -20,4 +21,57 @@ __u64 rust_helper_drm_vma_node_offset_addr(struct drm_vma_offset_node *node)
return drm_vma_node_offset_addr(node);
}
-#endif
+#if IS_ENABLED(CONFIG_DRM_GEM_SHMEM_HELPER)
+__rust_helper void
+rust_helper_drm_gem_shmem_object_free(struct drm_gem_object *obj)
+{
+ return drm_gem_shmem_object_free(obj);
+}
+
+__rust_helper void
+rust_helper_drm_gem_shmem_object_print_info(struct drm_printer *p, unsigned int indent,
+ const struct drm_gem_object *obj)
+{
+ drm_gem_shmem_object_print_info(p, indent, obj);
+}
+
+__rust_helper int
+rust_helper_drm_gem_shmem_object_pin(struct drm_gem_object *obj)
+{
+ return drm_gem_shmem_object_pin(obj);
+}
+
+__rust_helper void
+rust_helper_drm_gem_shmem_object_unpin(struct drm_gem_object *obj)
+{
+ drm_gem_shmem_object_unpin(obj);
+}
+
+__rust_helper struct sg_table *
+rust_helper_drm_gem_shmem_object_get_sg_table(struct drm_gem_object *obj)
+{
+ return drm_gem_shmem_object_get_sg_table(obj);
+}
+
+__rust_helper int
+rust_helper_drm_gem_shmem_object_vmap(struct drm_gem_object *obj,
+ struct iosys_map *map)
+{
+ return drm_gem_shmem_object_vmap(obj, map);
+}
+
+__rust_helper void
+rust_helper_drm_gem_shmem_object_vunmap(struct drm_gem_object *obj,
+ struct iosys_map *map)
+{
+ drm_gem_shmem_object_vunmap(obj, map);
+}
+
+__rust_helper int
+rust_helper_drm_gem_shmem_object_mmap(struct drm_gem_object *obj, struct vm_area_struct *vma)
+{
+ return drm_gem_shmem_object_mmap(obj, vma);
+}
+
+#endif /* CONFIG_DRM_GEM_SHMEM_HELPER */
+#endif /* CONFIG_DRM */
diff --git a/rust/kernel/drm/gem/mod.rs b/rust/kernel/drm/gem/mod.rs
index 972d50d4342dd..379ae3dfb02f5 100644
--- a/rust/kernel/drm/gem/mod.rs
+++ b/rust/kernel/drm/gem/mod.rs
@@ -3,6 +3,8 @@
//! DRM GEM API
//!
//! C header: [`include/drm/drm_gem.h`](srctree/include/drm/drm_gem.h)
+#[cfg(CONFIG_DRM_GEM_SHMEM_HELPER)]
+pub mod shmem;
use crate::{
alloc::flags::*,
@@ -50,6 +52,8 @@ unsafe fn dec_ref(obj: core::ptr::NonNull<Self>) {
};
}
+pub(crate) use impl_aref_for_gem_obj;
+
/// A type alias for retrieving a [`Driver`]s [`DriverFile`] implementation from its
/// [`DriverObject`] implementation.
///
@@ -205,7 +209,6 @@ fn create_mmap_offset(&self) -> Result<u64> {
impl<T: IntoGEMObject> BaseObject for T {}
/// Crate-private base operations shared by all GEM object classes.
-#[expect(unused)]
pub(crate) trait BaseObjectPrivate: IntoGEMObject {
/// Return a pointer to this object's dma_resv.
fn raw_dma_resv(&self) -> *mut bindings::dma_resv {
diff --git a/rust/kernel/drm/gem/shmem.rs b/rust/kernel/drm/gem/shmem.rs
new file mode 100644
index 0000000000000..d9f1a4e95cedc
--- /dev/null
+++ b/rust/kernel/drm/gem/shmem.rs
@@ -0,0 +1,249 @@
+// SPDX-License-Identifier: GPL-2.0
+
+//! DRM GEM shmem helper objects
+//!
+//! C header: [`include/linux/drm/drm_gem_shmem_helper.h`](srctree/include/linux/drm/drm_gem_shmem_helper.h)
+
+// TODO:
+// - There are a number of spots here that manually acquire/release the DMA reservation lock using
+// dma_resv_(un)lock(). In the future we should add support for ww mutex, expose a method to
+// acquire a reference to the WwMutex, and then use that directly instead of the C functions here.
+
+use crate::{
+ container_of,
+ drm::{
+ device,
+ driver,
+ gem,
+ private::Sealed, //
+ },
+ error::{
+ from_err_ptr,
+ to_result, //
+ },
+ prelude::*,
+ scatterlist,
+ types::{
+ ARef,
+ Opaque, //
+ }, //
+};
+use core::{
+ ops::{
+ Deref,
+ DerefMut, //
+ },
+ ptr::NonNull,
+};
+use gem::{
+ BaseObjectPrivate,
+ DriverObject,
+ IntoGEMObject, //
+};
+
+/// A struct for controlling the creation of shmem-backed GEM objects.
+///
+/// This is used with [`Object::new()`] to control various properties that can only be set when
+/// initially creating a shmem-backed GEM object.
+#[derive(Default)]
+pub struct ObjectConfig<'a, T: DriverObject> {
+ /// Whether to set the write-combine map flag.
+ pub map_wc: bool,
+
+ /// Reuse the DMA reservation from another GEM object.
+ ///
+ /// The newly created [`Object`] will hold an owned refcount to `parent_resv_obj` if specified.
+ pub parent_resv_obj: Option<&'a Object<T>>,
+}
+
+/// A shmem-backed GEM object.
+///
+/// # Invariants
+///
+/// `obj` contains a valid initialized `struct drm_gem_shmem_object` for the lifetime of this
+/// object.
+#[repr(C)]
+#[pin_data]
+pub struct Object<T: DriverObject> {
+ #[pin]
+ obj: Opaque<bindings::drm_gem_shmem_object>,
+ // Parent object that owns this object's DMA reservation object
+ parent_resv_obj: Option<ARef<Object<T>>>,
+ #[pin]
+ inner: T,
+}
+
+super::impl_aref_for_gem_obj!(impl<T> for Object<T> where T: DriverObject);
+
+// SAFETY: All GEM objects are thread-safe.
+unsafe impl<T: DriverObject> Send for Object<T> {}
+
+// SAFETY: All GEM objects are thread-safe.
+unsafe impl<T: DriverObject> Sync for Object<T> {}
+
+impl<T: DriverObject> Object<T> {
+ /// `drm_gem_object_funcs` vtable suitable for GEM shmem objects.
+ const VTABLE: bindings::drm_gem_object_funcs = bindings::drm_gem_object_funcs {
+ free: Some(Self::free_callback),
+ open: Some(super::open_callback::<T>),
+ close: Some(super::close_callback::<T>),
+ print_info: Some(bindings::drm_gem_shmem_object_print_info),
+ export: None,
+ pin: Some(bindings::drm_gem_shmem_object_pin),
+ unpin: Some(bindings::drm_gem_shmem_object_unpin),
+ get_sg_table: Some(bindings::drm_gem_shmem_object_get_sg_table),
+ vmap: Some(bindings::drm_gem_shmem_object_vmap),
+ vunmap: Some(bindings::drm_gem_shmem_object_vunmap),
+ mmap: Some(bindings::drm_gem_shmem_object_mmap),
+ status: None,
+ rss: None,
+ // SAFETY: `drm_gem_shmem_vm_ops` is static const on the C side, so immutable references are
+ // safe here and such references shall be valid forever
+ vm_ops: unsafe { &bindings::drm_gem_shmem_vm_ops },
+ evict: None,
+ };
+
+ /// Return a raw pointer to the embedded drm_gem_shmem_object.
+ fn as_shmem(&self) -> *mut bindings::drm_gem_shmem_object {
+ self.obj.get()
+ }
+
+ /// Create a new shmem-backed DRM object of the given size.
+ ///
+ /// Additional config options can be specified using `config`.
+ pub fn new(
+ dev: &device::Device<T::Driver>,
+ size: usize,
+ config: ObjectConfig<'_, T>,
+ args: T::Args,
+ ) -> Result<ARef<Self>> {
+ let new: Pin<KBox<Self>> = KBox::try_pin_init(
+ try_pin_init!(Self {
+ obj <- Opaque::init_zeroed(),
+ parent_resv_obj: config.parent_resv_obj.map(|p| p.into()),
+ inner <- T::new(dev, size, args),
+ }),
+ GFP_KERNEL,
+ )?;
+
+ // SAFETY: `obj.as_raw()` is guaranteed to be valid by the initialization above.
+ unsafe { (*new.as_raw()).funcs = &Self::VTABLE };
+
+ // SAFETY: The arguments are all valid via the type invariants.
+ to_result(unsafe { bindings::drm_gem_shmem_init(dev.as_raw(), new.as_shmem(), size) })?;
+
+ // SAFETY: We never move out of `self`.
+ let new = KBox::into_raw(unsafe { Pin::into_inner_unchecked(new) });
+
+ // SAFETY: We're taking over the owned refcount from `drm_gem_shmem_init`.
+ let obj = unsafe { ARef::from_raw(NonNull::new_unchecked(new)) };
+
+ // Start filling out values from `config`
+ if let Some(parent_resv) = config.parent_resv_obj {
+ // SAFETY: We have yet to expose the new gem object outside of this function, so it is
+ // safe to modify this field.
+ unsafe { (*obj.obj.get()).base.resv = parent_resv.raw_dma_resv() };
+ }
+
+ // SAFETY: We have yet to expose this object outside of this function, so we're guaranteed
+ // to have exclusive access - thus making this safe to hold a mutable reference to.
+ let shmem = unsafe { &mut *obj.as_shmem() };
+ shmem.set_map_wc(config.map_wc);
+
+ Ok(obj)
+ }
+
+ /// Returns the `Device` that owns this GEM object.
+ pub fn dev(&self) -> &device::Device<T::Driver> {
+ // SAFETY: `dev` will have been initialized in `Self::new()` by `drm_gem_shmem_init()`.
+ unsafe { device::Device::from_raw((*self.as_raw()).dev) }
+ }
+
+ extern "C" fn free_callback(obj: *mut bindings::drm_gem_object) {
+ // SAFETY:
+ // - DRM always passes a valid gem object here
+ // - We used drm_gem_shmem_create() in our create_gem_object callback, so we know that
+ // `obj` is contained within a drm_gem_shmem_object
+ let this = unsafe { container_of!(obj, bindings::drm_gem_shmem_object, base) };
+
+ // SAFETY:
+ // - We're in free_callback - so this function is safe to call.
+ // - We won't be using the gem resources on `this` after this call.
+ unsafe { bindings::drm_gem_shmem_release(this) };
+
+ // SAFETY:
+ // - We verified above that `obj` is valid, which makes `this` valid
+ // - This function is set in AllocOps, so we know that `this` is contained within a
+ // `Object<T>`
+ let this = unsafe { container_of!(Opaque::cast_from(this), Self, obj) }.cast_mut();
+
+ // SAFETY: We're recovering the Kbox<> we created in gem_create_object()
+ let _ = unsafe { KBox::from_raw(this) };
+ }
+
+ /// Creates (if necessary) and returns an immutable reference to a scatter-gather table of DMA
+ /// pages for this object.
+ ///
+ /// This will pin the object in memory.
+ #[inline]
+ pub fn sg_table(&self) -> Result<&scatterlist::SGTable> {
+ // SAFETY:
+ // - drm_gem_shmem_get_pages_sgt is thread-safe.
+ // - drm_gem_shmem_get_pages_sgt returns either a valid pointer to a scatterlist, or an
+ // error pointer.
+ let sgt = from_err_ptr(unsafe { bindings::drm_gem_shmem_get_pages_sgt(self.as_shmem()) })?;
+
+ // SAFETY: We checked above that `sgt` is not an error pointer, so it must be a valid
+ // pointer to a scatterlist
+ Ok(unsafe { scatterlist::SGTable::from_raw(sgt) })
+ }
+}
+
+impl<T: DriverObject> Deref for Object<T> {
+ type Target = T;
+
+ fn deref(&self) -> &Self::Target {
+ &self.inner
+ }
+}
+
+impl<T: DriverObject> DerefMut for Object<T> {
+ fn deref_mut(&mut self) -> &mut Self::Target {
+ &mut self.inner
+ }
+}
+
+impl<T: DriverObject> Sealed for Object<T> {}
+
+impl<T: DriverObject> gem::IntoGEMObject for Object<T> {
+ fn as_raw(&self) -> *mut bindings::drm_gem_object {
+ // SAFETY:
+ // - Our immutable reference is proof that this is safe to dereference.
+ // - `obj` is always a valid drm_gem_shmem_object via our type invariants.
+ unsafe { &raw mut (*self.obj.get()).base }
+ }
+
+ unsafe fn from_raw<'a>(obj: *mut bindings::drm_gem_object) -> &'a Object<T> {
+ // SAFETY: The safety contract of from_gem_obj() guarantees that `obj` is contained within
+ // `Self`
+ unsafe {
+ let obj = Opaque::cast_from(container_of!(obj, bindings::drm_gem_shmem_object, base));
+
+ &*container_of!(obj, Object<T>, obj)
+ }
+ }
+}
+
+impl<T: DriverObject> driver::AllocImpl for Object<T> {
+ type Driver = T::Driver;
+
+ const ALLOC_OPS: driver::AllocOps = driver::AllocOps {
+ gem_create_object: None,
+ prime_handle_to_fd: None,
+ prime_fd_to_handle: None,
+ gem_prime_import: None,
+ gem_prime_import_sg_table: Some(bindings::drm_gem_shmem_prime_import_sg_table),
+ dumb_create: Some(bindings::drm_gem_shmem_dumb_create),
+ dumb_map_offset: None,
+ };
+}
--
2.53.0
^ permalink raw reply related [flat|nested] 16+ messages in thread* Re: [PATCH v7 4/7] rust: drm: gem: shmem: Add DRM shmem helper abstraction
2026-02-06 22:34 ` [PATCH v7 4/7] rust: drm: gem: shmem: Add DRM shmem helper abstraction Lyude Paul
@ 2026-02-17 19:38 ` Daniel Almeida
0 siblings, 0 replies; 16+ messages in thread
From: Daniel Almeida @ 2026-02-17 19:38 UTC (permalink / raw)
To: Lyude Paul
Cc: linux-kernel, dri-devel, rust-for-linux, Danilo Krummrich,
nouveau, Gary Guo, Benno Lossin, Alexandre Courbot, Janne Grunau
Hi Lyude,
> On 6 Feb 2026, at 19:34, Lyude Paul <lyude@redhat.com> wrote:
>
> From: Asahi Lina <lina@asahilina.net>
>
> The DRM shmem helper includes common code useful for drivers which
> allocate GEM objects as anonymous shmem. Add a Rust abstraction for
> this. Drivers can choose the raw GEM implementation or the shmem layer,
> depending on their needs.
>
> Signed-off-by: Asahi Lina <lina@asahilina.net>
> Signed-off-by: Daniel Almeida <daniel.almeida@collabora.com>
> Signed-off-by: Lyude Paul <lyude@redhat.com>
>
> ---
>
> V2:
> * Use the drm_gem_shmem_init() and drm_gem_shmem_release() that I extracted
> so we can handle memory allocation in rust, which means we no longer have
> to handle freeing rust members of the struct by hand and have a closer
> implementation to the main gem object
> (this also gets rid of gem_create_object)
> * Get rid of GemObjectRef and UniqueGemObjectRef, we have ARef<T> at home.
> * Use Device<T::Driver> in Object<T>
> * Cleanup Object::<T>::new() a bit:
> * Cleanup safety comment
> * Use cast_mut()
> * Just import container_of!(), we use it all over anyhow
> * mut_shmem() -> as_shmem(), make it safe (there's no reason for being unsafe)
> * Remove any *const and *muts in structs, just use NonNull
> * Get rid of the previously hand-rolled sg_table bindings in shmem, use the
> bindings from Abdiel's sg_table patch series
> * Add a TODO at the top about DMA reservation APIs and a desire for WwMutex
> * Get rid of map_wc() and replace it with a new ObjectConfig struct. While
> it currently only specifies the map_wc flag, the idea here is that
> settings like map_wc() and parent_resv_obj() shouldn't be exposed as
> normal functions since the only place where it's safe to set them is
> when we're still guaranteed unique access to the GEM object, e.g. before
> returning it to the caller. Using a struct instead of individual
> arguments here is mainly because we'll be adding at least one more
> argument, and there's enough other gem shmem settings that trying to add
> all of them as individual function arguments in the future would be a bit
> messy.
> * Get rid of vm_numa_fields!, Lina didn't like this macro much either and I
> think that it's fine for us to just specify the #[cfg(…)] attributes by
> hand since we only need to do it twice.
> * Set drm_gem_object_funcs.vm_ops directly to drm_gem_shmem_vm_ops, don't
> export the various shmem funcs. I'm not sure why this wasn't possible
> before but it seems to work fine now.
> V4:
> * Switch from OpaqueObject to a normal Object<T> now that we've removed it
> * Remove `dev` from Object, it's redundant as drm_gem_object already has a
> device pointer we can use.
> * Use &raw instead of addr_of!()
> V5:
> * Fix typo in shmem::Object::as_raw()
> * Add type invariant around `obj` always being a valid
> `drm_gem_shmem_object` for the duration of the lifetime of `Object`.
> * Use Opaque::cast_from() instead of unrestricted casts
> * Use IS_ENABLED() for gem shmem C helpers.
> V7:
> * Fix import style
> * Don't forget to make Object<T> Send/Sync
>
> rust/bindings/bindings_helper.h | 2 +
> rust/helpers/drm.c | 56 ++++++-
> rust/kernel/drm/gem/mod.rs | 5 +-
> rust/kernel/drm/gem/shmem.rs | 249 ++++++++++++++++++++++++++++++++
> 4 files changed, 310 insertions(+), 2 deletions(-)
> create mode 100644 rust/kernel/drm/gem/shmem.rs
>
> diff --git a/rust/bindings/bindings_helper.h b/rust/bindings/bindings_helper.h
> index ecf31681df806..c93afbd095e87 100644
> --- a/rust/bindings/bindings_helper.h
> +++ b/rust/bindings/bindings_helper.h
> @@ -33,6 +33,7 @@
> #include <drm/drm_drv.h>
> #include <drm/drm_file.h>
> #include <drm/drm_gem.h>
> +#include <drm/drm_gem_shmem_helper.h>
> #include <drm/drm_ioctl.h>
> #include <kunit/test.h>
> #include <linux/auxiliary_bus.h>
> @@ -61,6 +62,7 @@
> #include <linux/fs.h>
> #include <linux/i2c.h>
> #include <linux/ioport.h>
> +#include <linux/iosys-map.h>
> #include <linux/jiffies.h>
> #include <linux/jump_label.h>
> #include <linux/mdio.h>
> diff --git a/rust/helpers/drm.c b/rust/helpers/drm.c
> index 450b406c6f273..53ec06879db3d 100644
> --- a/rust/helpers/drm.c
> +++ b/rust/helpers/drm.c
> @@ -1,6 +1,7 @@
> // SPDX-License-Identifier: GPL-2.0
>
> #include <drm/drm_gem.h>
> +#include <drm/drm_gem_shmem_helper.h>
> #include <drm/drm_vma_manager.h>
>
> #ifdef CONFIG_DRM
> @@ -20,4 +21,57 @@ __u64 rust_helper_drm_vma_node_offset_addr(struct drm_vma_offset_node *node)
> return drm_vma_node_offset_addr(node);
> }
>
> -#endif
> +#if IS_ENABLED(CONFIG_DRM_GEM_SHMEM_HELPER)
> +__rust_helper void
> +rust_helper_drm_gem_shmem_object_free(struct drm_gem_object *obj)
> +{
> + return drm_gem_shmem_object_free(obj);
> +}
> +
> +__rust_helper void
> +rust_helper_drm_gem_shmem_object_print_info(struct drm_printer *p, unsigned int indent,
> + const struct drm_gem_object *obj)
> +{
> + drm_gem_shmem_object_print_info(p, indent, obj);
> +}
> +
> +__rust_helper int
> +rust_helper_drm_gem_shmem_object_pin(struct drm_gem_object *obj)
> +{
> + return drm_gem_shmem_object_pin(obj);
> +}
> +
> +__rust_helper void
> +rust_helper_drm_gem_shmem_object_unpin(struct drm_gem_object *obj)
> +{
> + drm_gem_shmem_object_unpin(obj);
> +}
> +
> +__rust_helper struct sg_table *
> +rust_helper_drm_gem_shmem_object_get_sg_table(struct drm_gem_object *obj)
> +{
> + return drm_gem_shmem_object_get_sg_table(obj);
> +}
> +
> +__rust_helper int
> +rust_helper_drm_gem_shmem_object_vmap(struct drm_gem_object *obj,
> + struct iosys_map *map)
> +{
> + return drm_gem_shmem_object_vmap(obj, map);
> +}
> +
> +__rust_helper void
> +rust_helper_drm_gem_shmem_object_vunmap(struct drm_gem_object *obj,
> + struct iosys_map *map)
> +{
> + drm_gem_shmem_object_vunmap(obj, map);
> +}
> +
> +__rust_helper int
> +rust_helper_drm_gem_shmem_object_mmap(struct drm_gem_object *obj, struct vm_area_struct *vma)
> +{
> + return drm_gem_shmem_object_mmap(obj, vma);
> +}
> +
> +#endif /* CONFIG_DRM_GEM_SHMEM_HELPER */
> +#endif /* CONFIG_DRM */
> diff --git a/rust/kernel/drm/gem/mod.rs b/rust/kernel/drm/gem/mod.rs
> index 972d50d4342dd..379ae3dfb02f5 100644
> --- a/rust/kernel/drm/gem/mod.rs
> +++ b/rust/kernel/drm/gem/mod.rs
> @@ -3,6 +3,8 @@
> //! DRM GEM API
> //!
> //! C header: [`include/drm/drm_gem.h`](srctree/include/drm/drm_gem.h)
> +#[cfg(CONFIG_DRM_GEM_SHMEM_HELPER)]
> +pub mod shmem;
>
> use crate::{
> alloc::flags::*,
> @@ -50,6 +52,8 @@ unsafe fn dec_ref(obj: core::ptr::NonNull<Self>) {
> };
> }
>
> +pub(crate) use impl_aref_for_gem_obj;
> +
> /// A type alias for retrieving a [`Driver`]s [`DriverFile`] implementation from its
> /// [`DriverObject`] implementation.
> ///
> @@ -205,7 +209,6 @@ fn create_mmap_offset(&self) -> Result<u64> {
> impl<T: IntoGEMObject> BaseObject for T {}
>
> /// Crate-private base operations shared by all GEM object classes.
> -#[expect(unused)]
> pub(crate) trait BaseObjectPrivate: IntoGEMObject {
> /// Return a pointer to this object's dma_resv.
> fn raw_dma_resv(&self) -> *mut bindings::dma_resv {
> diff --git a/rust/kernel/drm/gem/shmem.rs b/rust/kernel/drm/gem/shmem.rs
> new file mode 100644
> index 0000000000000..d9f1a4e95cedc
> --- /dev/null
> +++ b/rust/kernel/drm/gem/shmem.rs
> @@ -0,0 +1,249 @@
> +// SPDX-License-Identifier: GPL-2.0
> +
> +//! DRM GEM shmem helper objects
> +//!
> +//! C header: [`include/linux/drm/drm_gem_shmem_helper.h`](srctree/include/linux/drm/drm_gem_shmem_helper.h)
> +
> +// TODO:
> +// - There are a number of spots here that manually acquire/release the DMA reservation lock using
> +// dma_resv_(un)lock(). In the future we should add support for ww mutex, expose a method to
> +// acquire a reference to the WwMutex, and then use that directly instead of the C functions here.
Right, thanks to Onur we’re pretty close on getting ww_mutex I’d say.
> +
> +use crate::{
> + container_of,
> + drm::{
> + device,
> + driver,
> + gem,
> + private::Sealed, //
> + },
> + error::{
> + from_err_ptr,
> + to_result, //
> + },
> + prelude::*,
> + scatterlist,
> + types::{
> + ARef,
> + Opaque, //
> + }, //
> +};
> +use core::{
> + ops::{
> + Deref,
> + DerefMut, //
> + },
> + ptr::NonNull,
> +};
> +use gem::{
> + BaseObjectPrivate,
> + DriverObject,
> + IntoGEMObject, //
> +};
> +
> +/// A struct for controlling the creation of shmem-backed GEM objects.
> +///
> +/// This is used with [`Object::new()`] to control various properties that can only be set when
> +/// initially creating a shmem-backed GEM object.
> +#[derive(Default)]
> +pub struct ObjectConfig<'a, T: DriverObject> {
> + /// Whether to set the write-combine map flag.
> + pub map_wc: bool,
> +
> + /// Reuse the DMA reservation from another GEM object.
> + ///
> + /// The newly created [`Object`] will hold an owned refcount to `parent_resv_obj` if specified.
> + pub parent_resv_obj: Option<&'a Object<T>>,
> +}
> +
> +/// A shmem-backed GEM object.
> +///
> +/// # Invariants
> +///
> +/// `obj` contains a valid initialized `struct drm_gem_shmem_object` for the lifetime of this
> +/// object.
> +#[repr(C)]
> +#[pin_data]
> +pub struct Object<T: DriverObject> {
> + #[pin]
> + obj: Opaque<bindings::drm_gem_shmem_object>,
> + // Parent object that owns this object's DMA reservation object
> + parent_resv_obj: Option<ARef<Object<T>>>,
> + #[pin]
> + inner: T,
> +}
> +
> +super::impl_aref_for_gem_obj!(impl<T> for Object<T> where T: DriverObject);
> +
> +// SAFETY: All GEM objects are thread-safe.
> +unsafe impl<T: DriverObject> Send for Object<T> {}
> +
> +// SAFETY: All GEM objects are thread-safe.
> +unsafe impl<T: DriverObject> Sync for Object<T> {}
> +
> +impl<T: DriverObject> Object<T> {
> + /// `drm_gem_object_funcs` vtable suitable for GEM shmem objects.
> + const VTABLE: bindings::drm_gem_object_funcs = bindings::drm_gem_object_funcs {
> + free: Some(Self::free_callback),
> + open: Some(super::open_callback::<T>),
> + close: Some(super::close_callback::<T>),
> + print_info: Some(bindings::drm_gem_shmem_object_print_info),
> + export: None,
> + pin: Some(bindings::drm_gem_shmem_object_pin),
> + unpin: Some(bindings::drm_gem_shmem_object_unpin),
> + get_sg_table: Some(bindings::drm_gem_shmem_object_get_sg_table),
> + vmap: Some(bindings::drm_gem_shmem_object_vmap),
> + vunmap: Some(bindings::drm_gem_shmem_object_vunmap),
> + mmap: Some(bindings::drm_gem_shmem_object_mmap),
> + status: None,
> + rss: None,
> + // SAFETY: `drm_gem_shmem_vm_ops` is static const on the C side, so immutable references are
> + // safe here and such references shall be valid forever
> + vm_ops: unsafe { &bindings::drm_gem_shmem_vm_ops },
> + evict: None,
> + };
> +
> + /// Return a raw pointer to the embedded drm_gem_shmem_object.
> + fn as_shmem(&self) -> *mut bindings::drm_gem_shmem_object {
nit: I’d shoehorn the term “raw” here somehow because of the pointer. Like
as_raw_shmem() or something.
> + self.obj.get()
> + }
> +
> + /// Create a new shmem-backed DRM object of the given size.
> + ///
> + /// Additional config options can be specified using `config`.
> + pub fn new(
> + dev: &device::Device<T::Driver>,
> + size: usize,
> + config: ObjectConfig<'_, T>,
> + args: T::Args,
> + ) -> Result<ARef<Self>> {
> + let new: Pin<KBox<Self>> = KBox::try_pin_init(
> + try_pin_init!(Self {
> + obj <- Opaque::init_zeroed(),
> + parent_resv_obj: config.parent_resv_obj.map(|p| p.into()),
> + inner <- T::new(dev, size, args),
> + }),
> + GFP_KERNEL,
> + )?;
> +
> + // SAFETY: `obj.as_raw()` is guaranteed to be valid by the initialization above.
> + unsafe { (*new.as_raw()).funcs = &Self::VTABLE };
> +
> + // SAFETY: The arguments are all valid via the type invariants.
> + to_result(unsafe { bindings::drm_gem_shmem_init(dev.as_raw(), new.as_shmem(), size) })?;
> +
> + // SAFETY: We never move out of `self`.
> + let new = KBox::into_raw(unsafe { Pin::into_inner_unchecked(new) });
> +
> + // SAFETY: We're taking over the owned refcount from `drm_gem_shmem_init`.
> + let obj = unsafe { ARef::from_raw(NonNull::new_unchecked(new)) };
> +
> + // Start filling out values from `config`
> + if let Some(parent_resv) = config.parent_resv_obj {
> + // SAFETY: We have yet to expose the new gem object outside of this function, so it is
> + // safe to modify this field.
> + unsafe { (*obj.obj.get()).base.resv = parent_resv.raw_dma_resv() };
> + }
> +
> + // SAFETY: We have yet to expose this object outside of this function, so we're guaranteed
> + // to have exclusive access - thus making this safe to hold a mutable reference to.
> + let shmem = unsafe { &mut *obj.as_shmem() };
> + shmem.set_map_wc(config.map_wc);
> +
> + Ok(obj)
> + }
> +
> + /// Returns the `Device` that owns this GEM object.
> + pub fn dev(&self) -> &device::Device<T::Driver> {
> + // SAFETY: `dev` will have been initialized in `Self::new()` by `drm_gem_shmem_init()`.
> + unsafe { device::Device::from_raw((*self.as_raw()).dev) }
> + }
> +
> + extern "C" fn free_callback(obj: *mut bindings::drm_gem_object) {
> + // SAFETY:
> + // - DRM always passes a valid gem object here
> + // - We used drm_gem_shmem_create() in our create_gem_object callback, so we know that
> + // `obj` is contained within a drm_gem_shmem_object
> + let this = unsafe { container_of!(obj, bindings::drm_gem_shmem_object, base) };
> +
> + // SAFETY:
> + // - We're in free_callback - so this function is safe to call.
> + // - We won't be using the gem resources on `this` after this call.
> + unsafe { bindings::drm_gem_shmem_release(this) };
> +
> + // SAFETY:
> + // - We verified above that `obj` is valid, which makes `this` valid
> + // - This function is set in AllocOps, so we know that `this` is contained within a
> + // `Object<T>`
> + let this = unsafe { container_of!(Opaque::cast_from(this), Self, obj) }.cast_mut();
> +
> + // SAFETY: We're recovering the Kbox<> we created in gem_create_object()
> + let _ = unsafe { KBox::from_raw(this) };
> + }
> +
> + /// Creates (if necessary) and returns an immutable reference to a scatter-gather table of DMA
> + /// pages for this object.
> + ///
> + /// This will pin the object in memory.
> + #[inline]
> + pub fn sg_table(&self) -> Result<&scatterlist::SGTable> {
> + // SAFETY:
> + // - drm_gem_shmem_get_pages_sgt is thread-safe.
> + // - drm_gem_shmem_get_pages_sgt returns either a valid pointer to a scatterlist, or an
> + // error pointer.
> + let sgt = from_err_ptr(unsafe { bindings::drm_gem_shmem_get_pages_sgt(self.as_shmem()) })?;
> +
> + // SAFETY: We checked above that `sgt` is not an error pointer, so it must be a valid
> + // pointer to a scatterlist
> + Ok(unsafe { scatterlist::SGTable::from_raw(sgt) })
> + }
> +}
> +
> +impl<T: DriverObject> Deref for Object<T> {
> + type Target = T;
> +
> + fn deref(&self) -> &Self::Target {
> + &self.inner
> + }
> +}
> +
> +impl<T: DriverObject> DerefMut for Object<T> {
> + fn deref_mut(&mut self) -> &mut Self::Target {
> + &mut self.inner
> + }
> +}
> +
> +impl<T: DriverObject> Sealed for Object<T> {}
> +
> +impl<T: DriverObject> gem::IntoGEMObject for Object<T> {
> + fn as_raw(&self) -> *mut bindings::drm_gem_object {
> + // SAFETY:
> + // - Our immutable reference is proof that this is safe to dereference.
> + // - `obj` is always a valid drm_gem_shmem_object via our type invariants.
> + unsafe { &raw mut (*self.obj.get()).base }
> + }
> +
> + unsafe fn from_raw<'a>(obj: *mut bindings::drm_gem_object) -> &'a Object<T> {
> + // SAFETY: The safety contract of from_gem_obj() guarantees that `obj` is contained within
> + // `Self`
> + unsafe {
> + let obj = Opaque::cast_from(container_of!(obj, bindings::drm_gem_shmem_object, base));
> +
> + &*container_of!(obj, Object<T>, obj)
> + }
> + }
> +}
> +
> +impl<T: DriverObject> driver::AllocImpl for Object<T> {
> + type Driver = T::Driver;
> +
> + const ALLOC_OPS: driver::AllocOps = driver::AllocOps {
> + gem_create_object: None,
> + prime_handle_to_fd: None,
> + prime_fd_to_handle: None,
> + gem_prime_import: None,
> + gem_prime_import_sg_table: Some(bindings::drm_gem_shmem_prime_import_sg_table),
> + dumb_create: Some(bindings::drm_gem_shmem_dumb_create),
> + dumb_map_offset: None,
> + };
> +}
> --
> 2.53.0
>
Looks good to me.
Reviewed-by: Daniel Almeida <daniel.almeida@collabora.com>
^ permalink raw reply [flat|nested] 16+ messages in thread
* [PATCH v7 5/7] rust: drm: gem: Introduce shmem::SGTable
2026-02-06 22:34 [PATCH v7 0/7] Rust bindings for gem shmem + iosys_map Lyude Paul
` (3 preceding siblings ...)
2026-02-06 22:34 ` [PATCH v7 4/7] rust: drm: gem: shmem: Add DRM shmem helper abstraction Lyude Paul
@ 2026-02-06 22:34 ` Lyude Paul
2026-02-17 19:53 ` Daniel Almeida
2026-02-06 22:34 ` [PATCH v7 6/7] rust: Introduce iosys_map bindings Lyude Paul
2026-02-06 22:34 ` [PATCH v7 7/7] rust: drm/gem: Add vmap functions to shmem bindings Lyude Paul
6 siblings, 1 reply; 16+ messages in thread
From: Lyude Paul @ 2026-02-06 22:34 UTC (permalink / raw)
To: linux-kernel, dri-devel, rust-for-linux, Danilo Krummrich
Cc: nouveau, Daniel Almeida, Gary Guo, Benno Lossin,
Alexandre Courbot, Janne Grunau
Currently we expose the ability to retrieve an SGTable for an shmem gem
object using gem::shmem::Object::<T>::sg_table(). However, this only gives
us a borrowed reference. This being said - retrieving an SGTable is a
fallible operation, and as such it's reasonable that a driver may want to
hold onto an SGTable for longer then a reference would allow in order to
avoid having to deal with fallibility every time they want to access the
SGTable. One such driver with this usecase is the Asahi driver.
So to support this, let's introduce shmem::SGTable - which both holds a
pointer to the SGTable and a reference to its respective GEM object in
order to keep the GEM object alive for as long as the shmem::SGTable. The
type can be used identically to a normal SGTable.
Signed-off-by: Lyude Paul <lyude@redhat.com>
Reviewed-by: Janne Grunau <j@jananu.net>
---
V3:
* Rename OwnedSGTable to shmem::SGTable. Since the current version of the
SGTable abstractions now has a `Owned` and `Borrowed` variant, I think
renaming this to shmem::SGTable makes things less confusing.
We do however, keep the name of owned_sg_table() as-is.
V4:
* Clarify safety comments for SGTable to explain why the object is
thread-safe.
* Rename from SGTableRef to SGTable
rust/kernel/drm/gem/shmem.rs | 50 ++++++++++++++++++++++++++++++++++++
1 file changed, 50 insertions(+)
diff --git a/rust/kernel/drm/gem/shmem.rs b/rust/kernel/drm/gem/shmem.rs
index d9f1a4e95cedc..e511a9b6710e0 100644
--- a/rust/kernel/drm/gem/shmem.rs
+++ b/rust/kernel/drm/gem/shmem.rs
@@ -197,6 +197,25 @@ pub fn sg_table(&self) -> Result<&scatterlist::SGTable> {
// pointer to a scatterlist
Ok(unsafe { scatterlist::SGTable::from_raw(sgt) })
}
+
+ /// Creates (if necessary) and returns an owned reference to a scatter-gather table of DMA pages
+ /// for this object.
+ ///
+ /// This is the same as [`sg_table`](Self::sg_table), except that it instead returns an
+ /// [`shmem::SGTable`] which holds a reference to the associated gem object, instead of a
+ /// reference to an [`scatterlist::SGTable`].
+ ///
+ /// This will pin the object in memory.
+ ///
+ /// [`shmem::SGTable`]: SGTable
+ pub fn owned_sg_table(&self) -> Result<SGTable<T>> {
+ Ok(SGTable {
+ sgt: self.sg_table()?.into(),
+ // INVARIANT: We take an owned refcount to `self` here, ensuring that `sgt` remains
+ // valid for as long as this `SGTable`.
+ _owner: self.into(),
+ })
+ }
}
impl<T: DriverObject> Deref for Object<T> {
@@ -247,3 +266,34 @@ impl<T: DriverObject> driver::AllocImpl for Object<T> {
dumb_map_offset: None,
};
}
+
+/// An owned reference to a scatter-gather table of DMA address spans for a GEM shmem object.
+///
+/// This object holds an owned reference to the underlying GEM shmem object, ensuring that the
+/// [`scatterlist::SGTable`] referenced by this type remains valid for the lifetime of this object.
+///
+/// # Invariants
+///
+/// - `sgt` is kept alive by `_owner`, ensuring it remains valid for as long as `Self`.
+/// - `sgt` corresponds to the owned object in `_owner`.
+/// - This object is only exposed in situations where we know the underlying `SGTable` will not be
+/// modified for the lifetime of this object. Thus, it is safe to send/access this type across
+/// threads.
+pub struct SGTable<T: DriverObject> {
+ sgt: NonNull<scatterlist::SGTable>,
+ _owner: ARef<Object<T>>,
+}
+
+// SAFETY: This object is thread-safe via our type invariants.
+unsafe impl<T: DriverObject> Send for SGTable<T> {}
+// SAFETY: This object is thread-safe via our type invariants.
+unsafe impl<T: DriverObject> Sync for SGTable<T> {}
+
+impl<T: DriverObject> Deref for SGTable<T> {
+ type Target = scatterlist::SGTable;
+
+ fn deref(&self) -> &Self::Target {
+ // SAFETY: Creating an immutable reference to this is safe via our type invariants.
+ unsafe { self.sgt.as_ref() }
+ }
+}
--
2.53.0
^ permalink raw reply related [flat|nested] 16+ messages in thread* Re: [PATCH v7 5/7] rust: drm: gem: Introduce shmem::SGTable
2026-02-06 22:34 ` [PATCH v7 5/7] rust: drm: gem: Introduce shmem::SGTable Lyude Paul
@ 2026-02-17 19:53 ` Daniel Almeida
0 siblings, 0 replies; 16+ messages in thread
From: Daniel Almeida @ 2026-02-17 19:53 UTC (permalink / raw)
To: Lyude Paul
Cc: linux-kernel, dri-devel, rust-for-linux, Danilo Krummrich,
nouveau, Gary Guo, Benno Lossin, Alexandre Courbot, Janne Grunau
> On 6 Feb 2026, at 19:34, Lyude Paul <lyude@redhat.com> wrote:
>
> Currently we expose the ability to retrieve an SGTable for an shmem gem
> object using gem::shmem::Object::<T>::sg_table(). However, this only gives
> us a borrowed reference. This being said - retrieving an SGTable is a
> fallible operation, and as such it's reasonable that a driver may want to
> hold onto an SGTable for longer then a reference would allow in order to
> avoid having to deal with fallibility every time they want to access the
> SGTable. One such driver with this usecase is the Asahi driver.
>
> So to support this, let's introduce shmem::SGTable - which both holds a
> pointer to the SGTable and a reference to its respective GEM object in
> order to keep the GEM object alive for as long as the shmem::SGTable. The
> type can be used identically to a normal SGTable.
>
> Signed-off-by: Lyude Paul <lyude@redhat.com>
> Reviewed-by: Janne Grunau <j@jananu.net>
>
> ---
> V3:
> * Rename OwnedSGTable to shmem::SGTable. Since the current version of the
> SGTable abstractions now has a `Owned` and `Borrowed` variant, I think
> renaming this to shmem::SGTable makes things less confusing.
> We do however, keep the name of owned_sg_table() as-is.
> V4:
> * Clarify safety comments for SGTable to explain why the object is
> thread-safe.
> * Rename from SGTableRef to SGTable
>
> rust/kernel/drm/gem/shmem.rs | 50 ++++++++++++++++++++++++++++++++++++
> 1 file changed, 50 insertions(+)
>
> diff --git a/rust/kernel/drm/gem/shmem.rs b/rust/kernel/drm/gem/shmem.rs
> index d9f1a4e95cedc..e511a9b6710e0 100644
> --- a/rust/kernel/drm/gem/shmem.rs
> +++ b/rust/kernel/drm/gem/shmem.rs
> @@ -197,6 +197,25 @@ pub fn sg_table(&self) -> Result<&scatterlist::SGTable> {
> // pointer to a scatterlist
> Ok(unsafe { scatterlist::SGTable::from_raw(sgt) })
> }
> +
> + /// Creates (if necessary) and returns an owned reference to a scatter-gather table of DMA pages
> + /// for this object.
> + ///
> + /// This is the same as [`sg_table`](Self::sg_table), except that it instead returns an
> + /// [`shmem::SGTable`] which holds a reference to the associated gem object, instead of a
> + /// reference to an [`scatterlist::SGTable`].
> + ///
> + /// This will pin the object in memory.
> + ///
> + /// [`shmem::SGTable`]: SGTable
> + pub fn owned_sg_table(&self) -> Result<SGTable<T>> {
> + Ok(SGTable {
> + sgt: self.sg_table()?.into(),
> + // INVARIANT: We take an owned refcount to `self` here, ensuring that `sgt` remains
> + // valid for as long as this `SGTable`.
> + _owner: self.into(),
> + })
> + }
> }
>
> impl<T: DriverObject> Deref for Object<T> {
> @@ -247,3 +266,34 @@ impl<T: DriverObject> driver::AllocImpl for Object<T> {
> dumb_map_offset: None,
> };
> }
> +
> +/// An owned reference to a scatter-gather table of DMA address spans for a GEM shmem object.
> +///
> +/// This object holds an owned reference to the underlying GEM shmem object, ensuring that the
> +/// [`scatterlist::SGTable`] referenced by this type remains valid for the lifetime of this object.
> +///
> +/// # Invariants
> +///
> +/// - `sgt` is kept alive by `_owner`, ensuring it remains valid for as long as `Self`.
> +/// - `sgt` corresponds to the owned object in `_owner`.
> +/// - This object is only exposed in situations where we know the underlying `SGTable` will not be
> +/// modified for the lifetime of this object. Thus, it is safe to send/access this type across
> +/// threads.
> +pub struct SGTable<T: DriverObject> {
> + sgt: NonNull<scatterlist::SGTable>,
> + _owner: ARef<Object<T>>,
> +}
> +
> +// SAFETY: This object is thread-safe via our type invariants.
> +unsafe impl<T: DriverObject> Send for SGTable<T> {}
> +// SAFETY: This object is thread-safe via our type invariants.
> +unsafe impl<T: DriverObject> Sync for SGTable<T> {}
> +
> +impl<T: DriverObject> Deref for SGTable<T> {
> + type Target = scatterlist::SGTable;
> +
> + fn deref(&self) -> &Self::Target {
> + // SAFETY: Creating an immutable reference to this is safe via our type invariants.
> + unsafe { self.sgt.as_ref() }
> + }
> +}
> --
> 2.53.0
>
Reviewed-by: Daniel Almeida <daniel.almeida@collabora.com>
^ permalink raw reply [flat|nested] 16+ messages in thread
* [PATCH v7 6/7] rust: Introduce iosys_map bindings
2026-02-06 22:34 [PATCH v7 0/7] Rust bindings for gem shmem + iosys_map Lyude Paul
` (4 preceding siblings ...)
2026-02-06 22:34 ` [PATCH v7 5/7] rust: drm: gem: Introduce shmem::SGTable Lyude Paul
@ 2026-02-06 22:34 ` Lyude Paul
2026-02-09 11:17 ` Danilo Krummrich
2026-02-13 1:59 ` Deborah Brouwer
2026-02-06 22:34 ` [PATCH v7 7/7] rust: drm/gem: Add vmap functions to shmem bindings Lyude Paul
6 siblings, 2 replies; 16+ messages in thread
From: Lyude Paul @ 2026-02-06 22:34 UTC (permalink / raw)
To: linux-kernel, dri-devel, rust-for-linux, Danilo Krummrich
Cc: nouveau, Daniel Almeida, Gary Guo, Benno Lossin,
Alexandre Courbot, Janne Grunau
This introduces a set of bindings for working with iosys_map in rust code
using the new Io traits.
Signed-off-by: Lyude Paul <lyude@redhat.com>
---
V5:
- Fix incorrect field size being passed to iosys_map_memcpy_to()
- Add an additional unit test, basic_macro(), which can successfully catch
the above issue so it doesn't happen again in the future.
V6:
- Drop as_slice/as_mut_slice (Alice Rhyl)
V7:
- Start using Alexandre Courbot's wonderful Io, IoCapable and IoKnownSize
traits instead of trying to roll our own IO accessors. This also changes
the following:
- We don't have a custom AsBytes/FromBytes type that we carry around any
longer with maps
- We now have optional compile-time size checking
- We don't need our own unit tests anymore
- RawIoSysMap can be used for unsafe IO operations, because why not.
- IoSysMapRef::new() can fail now since it needs to ensure SIZE is valid.
- We don't implement Deref<RawIoSysMap> for IoSysMapRef any longer. The
main reason for this is that we want to avoid users having to manually
specify if they want the RawIoSysMap or IoSysMapRef variant functions
like io_read()/io_write().
This is fine IMHO, but to make sure things remain easy for APIs that
wrap around iosys map we make IoSysMapRef.raw_map pub(crate).
rust/helpers/helpers.c | 1 +
rust/helpers/iosys_map.c | 34 +++++++
rust/kernel/iosys_map.rs | 211 +++++++++++++++++++++++++++++++++++++++
rust/kernel/lib.rs | 1 +
4 files changed, 247 insertions(+)
create mode 100644 rust/helpers/iosys_map.c
create mode 100644 rust/kernel/iosys_map.rs
diff --git a/rust/helpers/helpers.c b/rust/helpers/helpers.c
index 1d3333cc0d2a8..bd8ad237aa02e 100644
--- a/rust/helpers/helpers.c
+++ b/rust/helpers/helpers.c
@@ -31,6 +31,7 @@
#include "irq.c"
#include "fs.c"
#include "io.c"
+#include "iosys_map.c"
#include "jump_label.c"
#include "kunit.c"
#include "maple_tree.c"
diff --git a/rust/helpers/iosys_map.c b/rust/helpers/iosys_map.c
new file mode 100644
index 0000000000000..6861d4ec48a4b
--- /dev/null
+++ b/rust/helpers/iosys_map.c
@@ -0,0 +1,34 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include <linux/iosys-map.h>
+#include <linux/types.h>
+
+#define rust_iosys_map_rd(type__) \
+ __rust_helper type__ \
+ rust_helper_iosys_map_rd_ ## type__(const struct iosys_map *map, size_t offset) \
+ { \
+ return iosys_map_rd(map, offset, type__); \
+ }
+#define rust_iosys_map_wr(type__) \
+ __rust_helper void \
+ rust_helper_iosys_map_wr_ ## type__(const struct iosys_map *map, size_t offset, \
+ type__ value) \
+ { \
+ iosys_map_wr(map, offset, type__, value); \
+ }
+
+rust_iosys_map_rd(u8);
+rust_iosys_map_rd(u16);
+rust_iosys_map_rd(u32);
+
+rust_iosys_map_wr(u8);
+rust_iosys_map_wr(u16);
+rust_iosys_map_wr(u32);
+
+#ifdef CONFIG_64BIT
+rust_iosys_map_rd(u64);
+rust_iosys_map_wr(u64);
+#endif
+
+#undef rust_iosys_map_rd
+#undef rust_iosys_map_wr
diff --git a/rust/kernel/iosys_map.rs b/rust/kernel/iosys_map.rs
new file mode 100644
index 0000000000000..2070f0fb42cb8
--- /dev/null
+++ b/rust/kernel/iosys_map.rs
@@ -0,0 +1,211 @@
+// SPDX-License-Identifier: GPL-2.0
+
+//! IO-agnostic memory mapping interfaces.
+//!
+//! This crate provides bindings for the `struct iosys_map` type, which provides a common interface
+//! for memory mappings which can reside within coherent memory, or within IO memory.
+//!
+//! C header: [`include/linux/iosys-map.h`](srctree/include/linux/pci.h)
+
+use crate::{
+ error::code::*,
+ io::{
+ Io,
+ IoCapable,
+ IoKnownSize, //
+ },
+ prelude::*, //
+};
+use bindings;
+use core::marker::PhantomData;
+
+/// Raw unsized representation of a `struct iosys_map`.
+///
+/// This struct is a transparent wrapper around `struct iosys_map`. The C API does not provide the
+/// size of the mapping by default, and thus this type also does not include the size of the
+/// mapping. As such, it cannot be used for actually accessing the underlying data pointed to by the
+/// mapping.
+///
+/// With the exception of kernel crates which may provide their own wrappers around `RawIoSysMap`,
+/// users will typically not interact with this type directly.
+#[repr(transparent)]
+pub struct RawIoSysMap<const SIZE: usize = 0>(bindings::iosys_map);
+
+impl<const SIZE: usize> RawIoSysMap<SIZE> {
+ /// Convert from a raw `bindings::iosys_map`.
+ #[expect(unused)]
+ #[inline]
+ pub(crate) fn from_raw(val: bindings::iosys_map) -> Self {
+ Self(val)
+ }
+
+ /// Returns whether the mapping is within IO memory space or not.
+ #[inline]
+ pub fn is_iomem(&self) -> bool {
+ self.0.is_iomem
+ }
+
+ /// Convert from a `RawIoSysMap<SIZE>` to a raw `bindings::iosys_map` ref.
+ #[expect(unused)]
+ #[inline]
+ pub(crate) fn as_raw(&self) -> &bindings::iosys_map {
+ &self.0
+ }
+
+ /// Convert from a `RawIoSysMap<SIZE>` to a raw mutable `bindings::iosys_map` ref.
+ #[allow(unused)]
+ #[inline]
+ pub(crate) fn as_raw_mut(&mut self) -> &mut bindings::iosys_map {
+ &mut self.0
+ }
+
+ /// Returns the address pointed to by this iosys map.
+ ///
+ /// Note that this address is not guaranteed to be valid, and may or may not reside in I/O
+ /// memory.
+ #[inline]
+ pub fn addr(&self) -> usize {
+ (if self.is_iomem() {
+ // SAFETY: We confirmed above that this iosys map is contained within iomem, so it's
+ // safe to read vaddr_iomem
+ unsafe { self.0.__bindgen_anon_1.vaddr_iomem }
+ } else {
+ // SAFETY: We confirmed above that this iosys map is not contaned within iomem, so it's
+ // safe to read vaddr.
+ unsafe { self.0.__bindgen_anon_1.vaddr }
+ }) as usize
+ }
+}
+
+// SAFETY: As we make no guarantees about the validity of the mapping, there's no issue with sending
+// this type between threads.
+unsafe impl<const SIZE: usize> Send for RawIoSysMap<SIZE> {}
+
+impl<const SIZE: usize> Clone for RawIoSysMap<SIZE> {
+ fn clone(&self) -> Self {
+ Self(self.0)
+ }
+}
+
+macro_rules! impl_raw_iosys_map_io_capable {
+ ($ty:ty, $read_fn:ident, $write_fn:ident) => {
+ impl<const SIZE: usize> IoCapable<$ty> for RawIoSysMap<SIZE> {
+ #[inline(always)]
+ unsafe fn io_read(&self, address: usize) -> $ty {
+ // SAFETY: By the trait invariant `address` is a valid address for iosys map
+ // operations.
+ unsafe { bindings::$read_fn(&self.0, address) }
+ }
+
+ #[inline(always)]
+ unsafe fn io_write(&self, value: $ty, address: usize) {
+ // SAFETY: By the trait invariant `address` is a valid address for iosys map
+ // operations.
+ unsafe { bindings::$write_fn(&self.0, address, value) };
+ }
+ }
+ };
+}
+
+impl_raw_iosys_map_io_capable!(u8, iosys_map_rd_u8, iosys_map_wr_u8);
+impl_raw_iosys_map_io_capable!(u16, iosys_map_rd_u16, iosys_map_wr_u16);
+impl_raw_iosys_map_io_capable!(u32, iosys_map_rd_u32, iosys_map_wr_u32);
+#[cfg(CONFIG_64BIT)]
+impl_raw_iosys_map_io_capable!(u64, iosys_map_rd_u64, iosys_map_wr_u64);
+
+/// A sized version of a [`RawIoSysMap`].
+///
+/// This type includes the runtime size of the [`RawIoSysMap`] and can be used for checked I/O
+/// operations.
+///
+/// # Invariants
+///
+/// - The iosys mapping referenced by this type is guaranteed to be of at least `size` bytes in
+/// size
+/// - The iosys mapping referenced by this type is valid for the lifetime `'a`.
+#[derive(Clone)]
+pub struct IoSysMapRef<'a, const SIZE: usize = 0> {
+ pub(crate) raw_map: RawIoSysMap<SIZE>,
+ size: usize,
+ _p: PhantomData<&'a ()>,
+}
+
+impl<'a, const SIZE: usize> IoSysMapRef<'a, SIZE> {
+ /// Create a new [`IoSysMapRef`] from a [`RawIoSysMap`].
+ ///
+ /// Returns an error if the specified size is invalid.
+ ///
+ /// # Safety
+ ///
+ /// - The caller guarantees that the mapping is of at least `size` bytes.
+ /// - The caller guarantees that the mapping remains valid for the lifetime of `'a`.
+ #[expect(unused)]
+ pub(crate) unsafe fn new(map: RawIoSysMap<SIZE>, size: usize) -> Result<Self> {
+ if size < SIZE {
+ return Err(EINVAL);
+ }
+
+ Ok(Self {
+ raw_map: map,
+ size,
+ _p: PhantomData,
+ })
+ }
+
+ /// Returns whether the mapping is within IO memory space or not.
+ #[inline]
+ pub fn is_iomem(&self) -> bool {
+ self.raw_map.is_iomem()
+ }
+
+ /// Returns the address pointed to by this iosys map.
+ ///
+ /// Note that this address is not guaranteed to be valid, and may or may not reside in I/O
+ /// memory.
+ #[inline]
+ pub fn addr(&self) -> usize {
+ self.raw_map.addr()
+ }
+}
+
+macro_rules! impl_iosys_map_io_capable {
+ ($ty:ty) => {
+ impl<'a, const SIZE: usize> IoCapable<$ty> for IoSysMapRef<'a, SIZE> {
+ #[inline(always)]
+ unsafe fn io_read(&self, address: usize) -> $ty {
+ // SAFETY: By the trait invariant `address` is a valid address for iosys map
+ // operations.
+ unsafe { self.raw_map.io_read(address) }
+ }
+
+ #[inline(always)]
+ unsafe fn io_write(&self, value: $ty, address: usize) {
+ // SAFETY: By the trait invariant `address` is a valid address for iosys map
+ // operations.
+ unsafe { self.raw_map.io_write(value, address) };
+ }
+ }
+ };
+}
+
+impl_iosys_map_io_capable!(u8);
+impl_iosys_map_io_capable!(u16);
+impl_iosys_map_io_capable!(u32);
+#[cfg(CONFIG_64BIT)]
+impl_iosys_map_io_capable!(u64);
+
+impl<'a, const SIZE: usize> Io for IoSysMapRef<'a, SIZE> {
+ #[inline]
+ fn addr(&self) -> usize {
+ self.raw_map.addr()
+ }
+
+ #[inline]
+ fn maxsize(&self) -> usize {
+ self.size
+ }
+}
+
+impl<'a, const SIZE: usize> IoKnownSize for IoSysMapRef<'a, SIZE> {
+ const MIN_SIZE: usize = SIZE;
+}
diff --git a/rust/kernel/lib.rs b/rust/kernel/lib.rs
index 6d637e2fed1b6..02385af66fde2 100644
--- a/rust/kernel/lib.rs
+++ b/rust/kernel/lib.rs
@@ -103,6 +103,7 @@
pub mod init;
pub mod io;
pub mod ioctl;
+pub mod iosys_map;
pub mod iov;
pub mod irq;
pub mod jump_label;
--
2.53.0
^ permalink raw reply related [flat|nested] 16+ messages in thread* Re: [PATCH v7 6/7] rust: Introduce iosys_map bindings
2026-02-06 22:34 ` [PATCH v7 6/7] rust: Introduce iosys_map bindings Lyude Paul
@ 2026-02-09 11:17 ` Danilo Krummrich
2026-02-13 1:59 ` Deborah Brouwer
1 sibling, 0 replies; 16+ messages in thread
From: Danilo Krummrich @ 2026-02-09 11:17 UTC (permalink / raw)
To: Lyude Paul
Cc: linux-kernel, dri-devel, rust-for-linux, nouveau, Daniel Almeida,
Gary Guo, Benno Lossin, Alexandre Courbot, Janne Grunau
On Fri Feb 6, 2026 at 11:34 PM CET, Lyude Paul wrote:
> +/// Raw unsized representation of a `struct iosys_map`.
> +///
> +/// This struct is a transparent wrapper around `struct iosys_map`. The C API does not provide the
> +/// size of the mapping by default, and thus this type also does not include the size of the
> +/// mapping. As such, it cannot be used for actually accessing the underlying data pointed to by the
> +/// mapping.
> +///
> +/// With the exception of kernel crates which may provide their own wrappers around `RawIoSysMap`,
> +/// users will typically not interact with this type directly.
> +#[repr(transparent)]
> +pub struct RawIoSysMap<const SIZE: usize = 0>(bindings::iosys_map);
I'm still against using struct iosys_map as a common frontend for I/O memory and
system memory.
Exposing this as another I/O backend instead of just having a Rust structure as
frontend for a "real" abstraction around the Rust backends has various
downsides.
(1) We are limited to the features of struct iosys_map. The corresponding Rust
backends may provide additional functionality, which we can't access with
struct iosys_map. For instance, they Mmio will provide a relaxed() method
providing access to a borrowed backend providing relaxed ordering
accessors.
(2) We loose out on the capability to consider compile time checks regarding
the guaranteed minimum size of the mapping. (To be fair this could be
implemented on `IoSysMap` itself as well, but it would duplicate code that
we already have in the corresponding backends.)
(3) You have to duplicate the safety requirements of the backends that struct
iosys_map wraps. In fact, this series ignores that if the backend is I/O
memory we have to guarantee the it is revoked when the device this I/O
memory originates from is unbound.
Having a look at patch 7, it should be possible to read `is_iomem` and `vaddr` /
`vaddr_iomem` from the struct iosys_map and just construct the "real" `Mmio`
backend from it. We also have to create a backend for normal system memory, but
that should be trivial. :)
^ permalink raw reply [flat|nested] 16+ messages in thread* Re: [PATCH v7 6/7] rust: Introduce iosys_map bindings
2026-02-06 22:34 ` [PATCH v7 6/7] rust: Introduce iosys_map bindings Lyude Paul
2026-02-09 11:17 ` Danilo Krummrich
@ 2026-02-13 1:59 ` Deborah Brouwer
2026-03-04 22:59 ` lyude
1 sibling, 1 reply; 16+ messages in thread
From: Deborah Brouwer @ 2026-02-13 1:59 UTC (permalink / raw)
To: Lyude Paul
Cc: linux-kernel, dri-devel, rust-for-linux, Danilo Krummrich,
nouveau, Daniel Almeida, Gary Guo, Benno Lossin,
Alexandre Courbot, Janne Grunau
Hi Lyude,
On Fri, Feb 06, 2026 at 05:34:30PM -0500, Lyude Paul wrote:
> This introduces a set of bindings for working with iosys_map in rust code
> using the new Io traits.
>
> Signed-off-by: Lyude Paul <lyude@redhat.com>
>
> ---
> V5:
> - Fix incorrect field size being passed to iosys_map_memcpy_to()
> - Add an additional unit test, basic_macro(), which can successfully catch
> the above issue so it doesn't happen again in the future.
> V6:
> - Drop as_slice/as_mut_slice (Alice Rhyl)
> V7:
> - Start using Alexandre Courbot's wonderful Io, IoCapable and IoKnownSize
> traits instead of trying to roll our own IO accessors. This also changes
> the following:
> - We don't have a custom AsBytes/FromBytes type that we carry around any
> longer with maps
> - We now have optional compile-time size checking
> - We don't need our own unit tests anymore
> - RawIoSysMap can be used for unsafe IO operations, because why not.
> - IoSysMapRef::new() can fail now since it needs to ensure SIZE is valid.
> - We don't implement Deref<RawIoSysMap> for IoSysMapRef any longer. The
> main reason for this is that we want to avoid users having to manually
> specify if they want the RawIoSysMap or IoSysMapRef variant functions
> like io_read()/io_write().
> This is fine IMHO, but to make sure things remain easy for APIs that
> wrap around iosys map we make IoSysMapRef.raw_map pub(crate).
>
> rust/helpers/helpers.c | 1 +
> rust/helpers/iosys_map.c | 34 +++++++
> rust/kernel/iosys_map.rs | 211 +++++++++++++++++++++++++++++++++++++++
> rust/kernel/lib.rs | 1 +
> 4 files changed, 247 insertions(+)
> create mode 100644 rust/helpers/iosys_map.c
> create mode 100644 rust/kernel/iosys_map.rs
>
> diff --git a/rust/helpers/helpers.c b/rust/helpers/helpers.c
> index 1d3333cc0d2a8..bd8ad237aa02e 100644
> --- a/rust/helpers/helpers.c
> +++ b/rust/helpers/helpers.c
> @@ -31,6 +31,7 @@
> #include "irq.c"
> #include "fs.c"
> #include "io.c"
> +#include "iosys_map.c"
> #include "jump_label.c"
> #include "kunit.c"
> #include "maple_tree.c"
> diff --git a/rust/helpers/iosys_map.c b/rust/helpers/iosys_map.c
> new file mode 100644
> index 0000000000000..6861d4ec48a4b
> --- /dev/null
> +++ b/rust/helpers/iosys_map.c
> @@ -0,0 +1,34 @@
> +// SPDX-License-Identifier: GPL-2.0
> +
> +#include <linux/iosys-map.h>
> +#include <linux/types.h>
> +
> +#define rust_iosys_map_rd(type__) \
> + __rust_helper type__ \
> + rust_helper_iosys_map_rd_ ## type__(const struct iosys_map *map, size_t offset) \
> + { \
> + return iosys_map_rd(map, offset, type__); \
> + }
> +#define rust_iosys_map_wr(type__) \
> + __rust_helper void \
> + rust_helper_iosys_map_wr_ ## type__(const struct iosys_map *map, size_t offset, \
> + type__ value) \
> + { \
> + iosys_map_wr(map, offset, type__, value); \
> + }
> +
> +rust_iosys_map_rd(u8);
> +rust_iosys_map_rd(u16);
> +rust_iosys_map_rd(u32);
> +
> +rust_iosys_map_wr(u8);
> +rust_iosys_map_wr(u16);
> +rust_iosys_map_wr(u32);
> +
> +#ifdef CONFIG_64BIT
> +rust_iosys_map_rd(u64);
> +rust_iosys_map_wr(u64);
> +#endif
> +
> +#undef rust_iosys_map_rd
> +#undef rust_iosys_map_wr
> diff --git a/rust/kernel/iosys_map.rs b/rust/kernel/iosys_map.rs
> new file mode 100644
> index 0000000000000..2070f0fb42cb8
> --- /dev/null
> +++ b/rust/kernel/iosys_map.rs
> @@ -0,0 +1,211 @@
> +// SPDX-License-Identifier: GPL-2.0
> +
> +//! IO-agnostic memory mapping interfaces.
> +//!
> +//! This crate provides bindings for the `struct iosys_map` type, which provides a common interface
> +//! for memory mappings which can reside within coherent memory, or within IO memory.
> +//!
> +//! C header: [`include/linux/iosys-map.h`](srctree/include/linux/pci.h)
> +
> +use crate::{
> + error::code::*,
> + io::{
> + Io,
> + IoCapable,
> + IoKnownSize, //
> + },
> + prelude::*, //
> +};
> +use bindings;
> +use core::marker::PhantomData;
> +
> +/// Raw unsized representation of a `struct iosys_map`.
> +///
> +/// This struct is a transparent wrapper around `struct iosys_map`. The C API does not provide the
> +/// size of the mapping by default, and thus this type also does not include the size of the
> +/// mapping. As such, it cannot be used for actually accessing the underlying data pointed to by the
> +/// mapping.
> +///
> +/// With the exception of kernel crates which may provide their own wrappers around `RawIoSysMap`,
> +/// users will typically not interact with this type directly.
> +#[repr(transparent)]
> +pub struct RawIoSysMap<const SIZE: usize = 0>(bindings::iosys_map);
> +
> +impl<const SIZE: usize> RawIoSysMap<SIZE> {
> + /// Convert from a raw `bindings::iosys_map`.
> + #[expect(unused)]
> + #[inline]
> + pub(crate) fn from_raw(val: bindings::iosys_map) -> Self {
> + Self(val)
> + }
> +
> + /// Returns whether the mapping is within IO memory space or not.
> + #[inline]
> + pub fn is_iomem(&self) -> bool {
> + self.0.is_iomem
> + }
> +
> + /// Convert from a `RawIoSysMap<SIZE>` to a raw `bindings::iosys_map` ref.
> + #[expect(unused)]
> + #[inline]
> + pub(crate) fn as_raw(&self) -> &bindings::iosys_map {
> + &self.0
> + }
> +
> + /// Convert from a `RawIoSysMap<SIZE>` to a raw mutable `bindings::iosys_map` ref.
> + #[allow(unused)]
> + #[inline]
> + pub(crate) fn as_raw_mut(&mut self) -> &mut bindings::iosys_map {
> + &mut self.0
> + }
> +
> + /// Returns the address pointed to by this iosys map.
> + ///
> + /// Note that this address is not guaranteed to be valid, and may or may not reside in I/O
> + /// memory.
> + #[inline]
> + pub fn addr(&self) -> usize {
> + (if self.is_iomem() {
> + // SAFETY: We confirmed above that this iosys map is contained within iomem, so it's
> + // safe to read vaddr_iomem
> + unsafe { self.0.__bindgen_anon_1.vaddr_iomem }
> + } else {
> + // SAFETY: We confirmed above that this iosys map is not contaned within iomem, so it's
> + // safe to read vaddr.
> + unsafe { self.0.__bindgen_anon_1.vaddr }
> + }) as usize
> + }
> +}
> +
> +// SAFETY: As we make no guarantees about the validity of the mapping, there's no issue with sending
> +// this type between threads.
> +unsafe impl<const SIZE: usize> Send for RawIoSysMap<SIZE> {}
> +
> +impl<const SIZE: usize> Clone for RawIoSysMap<SIZE> {
> + fn clone(&self) -> Self {
> + Self(self.0)
> + }
> +}
> +
> +macro_rules! impl_raw_iosys_map_io_capable {
> + ($ty:ty, $read_fn:ident, $write_fn:ident) => {
> + impl<const SIZE: usize> IoCapable<$ty> for RawIoSysMap<SIZE> {
> + #[inline(always)]
> + unsafe fn io_read(&self, address: usize) -> $ty {
> + // SAFETY: By the trait invariant `address` is a valid address for iosys map
> + // operations.
> + unsafe { bindings::$read_fn(&self.0, address) }
> + }
> +
> + #[inline(always)]
> + unsafe fn io_write(&self, value: $ty, address: usize) {
> + // SAFETY: By the trait invariant `address` is a valid address for iosys map
> + // operations.
> + unsafe { bindings::$write_fn(&self.0, address, value) };
> + }
> + }
> + };
> +}
> +
I think there might be a mismatch between the absolute address being
passed to these read and write functions and the bindings helpers
that expect an offset argument.
This crashed Tyr when I tried to write the firmware in u8 chunks at
incremental offsets. But if i just calculated the offset and passed that
instead of the absolute address, this worked fine:
let offset = address - self.addr();
unsafe { bindings::$write_fn(&self.0, offset, value) };
Here's some of the call trace:
[ 31.553727] tyr fb000000.gpu: supply sram not found, using dummy regulator
[ 31.555096] tyr fb000000.gpu: mali-unknown id 0xa867 major 0x67 minor 0x0 status 0x5
[ 31.555778] tyr fb000000.gpu: Features: L2:0x7120306 Tiler:0x809 Mem:0x301 MMU:0x2830 AS:0xff
[ 31.556521] tyr fb000000.gpu: shader_present=0x0000000000050005 l2_present=0x0000000000000001 tiler_present=0x0000000000000001
[ 31.557868] stackdepot: allocating hash table of 524288 entries via kvcalloc
[ OK ] Started systemd-tmpfiles-clean.tim…y Cleanup of Temporary Directories.
[ 31.562424] stackdepot: allocating space for 8192 stack pools via kvcalloc
[ OK ] Reached target timers.target - Timer Units.
[ 31.563676] tyr: Firmware protected mode entry not supported, ignoring
[ 31.571391] Unable to handle kernel paging request at virtual address 0000000000800069
[ 31.572762] Mem abort info:
[ 31.573027] ESR = 0x0000000096000021
[ 31.573402] EC = 0x25: DABT (current EL), IL = 32 bits
[ 31.573878] SET = 0, FnV = 0
[ 31.574157] EA = 0, S1PTW = 0
[ 31.574442] FSC = 0x21: alignment fault
[ OK ] Listening on dbus.socket - D-Bus System Message Bus Socket.
[ 31.593348] Data abort info:
[ 31.593628] ISV = 0, ISS = 0x00000021, ISS2 = 0x00000000
[ 31.594117] CM = 0, WnR = 0, TnD = 0, TagAccess = 0
[ 31.594567] GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0
[ 31.595042] user pgtable: 4k pages, 48-bit VAs, pgdp=0000000105e50000
[ 31.595613] [0000000000800069] pgd=0000000000000000, p4d=0000000000000000
[ 31.596434] Internal error: Oops: 0000000096000021 [#1] SMP
[ 31.596936] Modules linked in: snd tyr(+) soundcore sha256 cfg80211 rfkill pci_endpoint_test fuse dm_mod ipv6
[ 31.597830] CPU: 3 UID: 0 PID: 241 Comm: (udev-worker) Not tainted 6.19.0-rc5 #282 PREEMPT
[ 31.598561] Hardware name: Radxa ROCK 5B (DT)
[ 31.598944] pstate: 00400009 (nzcv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[ 31.599554] pc : __d_lookup_rcu+0xbc/0x168
[ 31.599921] lr : __d_lookup_rcu+0x60/0x168
[ 31.600283] sp : ffff800081ebba00
[ 31.600574] x29: ffff800081ebba00 x28: 0000000000000049 x27: ffff00010d07002b
[ 31.601205] x26: 0000000000000081 x25: 000000000006915e x24: ffff0002f6600000
[ 31.601848] x23: ffff00010396a040 x22: ffff000102098460 x21: 000000036915e207
[ 31.601859] x20: ffff000101e10af8 x19: ffff800081ebbcac x18: 2f64662f666c6573
[ 31.601867] x17: ffffff00666c6573 x16: 00000000fffffffc x15: 0000000000003431
[ 31.601875] x14: 000000000000000f x13: 0080205100800107 x12: 0000000000800069
[ 31.601882] x11: ffffffffffffffff x10: 0000000000000018 x9 : 0000000000000003
[ 31.601890] x8 : 00000000008000a1 x7 : ffffb9d13df173c8 x6 : 0000000000000000
[ 31.601897] x5 : 0000000000000000 x4 : 0000000000000010 x3 : ffffffffffff0a00
[ 31.601905] x2 : ffff800081ebbcac x1 : ffffb9d1406be718 x0 : 0000000000000001
[ 31.601913] Call trace:
[ 31.601915] __d_lookup_rcu+0xbc/0x168 (P)
[ 31.601922] lookup_fast+0x98/0x174
[ 31.601929] link_path_walk+0x280/0x528
[ 31.601935] path_lookupat+0x60/0x1f0
[ 31.601941] do_o_path+0x34/0xb4
[ 31.601947] path_openat+0xccc/0xe34
[ 31.601953] do_filp_open+0xc0/0x170
[ 31.601958] do_sys_openat2+0x88/0x10c
[ 31.601963] __arm64_sys_openat+0x70/0x9c
[ 31.601968] invoke_syscall+0x40/0xcc
[ 31.601974] el0_svc_common+0x80/0xd8
[ 31.601979] do_el0_svc+0x1c/0x28
[ 31.601984] el0_svc+0x54/0x1d4
[ 31.601991] el0t_64_sync_handler+0x84/0x12c
[ 31.601997] el0t_64_sync+0x198/0x19c
[ 31.602005] Code: 14000003 f9400108 b4000428 d100e10c (88dffd8c)
[ 31.602009] ---[ end trace 0000000000000000 ]---
[ OK ] Listening on sshd-unix-local.socke…temd-ssh-generator, AF_UNIX Local).
[ 31.620067] mc: Linux media interface: v0.10
[ 31.623483] [drm] Initialized panthor 1.5.0 for fb000000.gpu on minor 0
[ 31.623765] Unable to handle kernel paging request at virtual address 00802a4d0080284d
[ 31.624285] tyr fb000000.gpu: Tyr initialized correctly.
[ 31.624752] Mem abort info:
[ 31.624754] ESR = 0x0000000096000004
[ 31.625847] EC = 0x25: DABT (current EL), IL = 32 bits
[ 31.626310] SET = 0, FnV = 0
[ 31.626578] EA = 0, S1PTW = 0
[ 31.626853] FSC = 0x04: level 0 translation fault
[ 31.627277] Data abort info:
[ 31.627529] ISV = 0, ISS = 0x00000004, ISS2 = 0x00000000
[ 31.628006] CM = 0, WnR = 0, TnD = 0, TagAccess = 0
[ 31.628447] GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0
[ 31.628910] [00802a4d0080284d] address between user and kernel address ranges
[ 31.629535] Internal error: Oops: 0000000096000004 [#2] SMP
[ 31.630030] Modules linked in: mc drm_client_lib snd tyr soundcore sha256 cfg80211 rfkill pci_endpoint_test fuse dm_mod ipv6
[ 31.631017] CPU: 4 UID: 0 PID: 225 Comm: systemd-udevd Tainted: G D 6.19.0-rc5 #282 PREEMPT
[ 31.631877] Tainted: [D]=DIE
[ 31.632129] Hardware name: Radxa ROCK 5B (DT)
[ 31.632506] pstate: a0400009 (NzCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[ 31.633111] pc : ___d_drop+0xd8/0x144
[ 31.633433] lr : d_invalidate+0x3c/0x110
[ 31.633776] sp : ffff800081d3ba70
[ 31.634064] x29: ffff800081d3ba90 x28: 0000000000000001 x27: ffffb9d140bfa000
[ 31.634685] x26: ffff0001039b5108 x25: ffff00010396a000 x24: ffffb9d140166275
[ 31.635305] x23: ffffb9d140bfa000 x22: ffffb9d140152f3a x21: ffff000104076000
[ 31.635925] x20: ffff00010e520098 x19: ffff00010396a000 x18: 0000000000000000
[ 31.636545] x17: 0000000000000000 x16: 0000000000000000 x15: 0000000000000000
[ 31.637165] x14: 0000000000000000 x13: ffff00010792b8f0 x12: 000000000000017a
[ 31.637785] x11: 00000000008000a1 x10: 00802a4d0080284d x9 : ffff0002f6604010
[ 31.638405] x8 : ffff000105bfcd40 x7 : 0000000000000000 x6 : ffffb9d13e272c60
[ 31.639026] x5 : 0000000000000000 x4 : 0000000000000001 x3 : 0000000000000000
[ 31.639646] x2 : 0000000000000000 x1 : 0000000000000000 x0 : ffff00010396a000
[ 31.640266] Call trace:
[ 31.640479] ___d_drop+0xd8/0x144 (P)
[ 31.640800] d_invalidate+0x3c/0x110
[ 31.641113] proc_invalidate_siblings_dcache+0x244/0x2b8
[ 31.641577] proc_flush_pid+0x1c/0x28
[ 31.641897] release_task+0x560/0x668
[ 31.642218] wait_consider_task+0x5a0/0xb44
[ 31.642582] __do_wait+0x154/0x1f0
[ 31.642879] do_wait+0x84/0x16c
[ 31.643154] __arm64_sys_waitid+0xac/0x218
[ 31.643512] invoke_syscall+0x40/0xcc
[ 31.643833] el0_svc_common+0x80/0xd8
[ 31.644152] do_el0_svc+0x1c/0x28
[ 31.644441] el0_svc+0x54/0x1d4
[ 31.644718] el0t_64_sync_handler+0x84/0x12c
[ 31.645090] el0t_64_sync+0x198/0x19c
[ 31.645412] Code: 5280002a f85f83a0 17fffff1 a944280b (f940014c)
[ 31.645941] ---[ end trace 0000000000000000 ]---
thanks,
Deborah
^ permalink raw reply [flat|nested] 16+ messages in thread* Re: [PATCH v7 6/7] rust: Introduce iosys_map bindings
2026-02-13 1:59 ` Deborah Brouwer
@ 2026-03-04 22:59 ` lyude
2026-03-06 20:59 ` Deborah Brouwer
0 siblings, 1 reply; 16+ messages in thread
From: lyude @ 2026-03-04 22:59 UTC (permalink / raw)
To: Deborah Brouwer
Cc: linux-kernel, dri-devel, rust-for-linux, Danilo Krummrich,
nouveau, Daniel Almeida, Gary Guo, Benno Lossin,
Alexandre Courbot, Janne Grunau
Hey! Sorry for the long delay in response
Would you mind sending me the size/offset numbers you used here? My
unit tests should have caught issues like this, so if they're happening
I definitely should add it to the testcases I've got to make sure
there's no problems
On Thu, 2026-02-12 at 17:59 -0800, Deborah Brouwer wrote:
> Hi Lyude,
>
> On Fri, Feb 06, 2026 at 05:34:30PM -0500, Lyude Paul wrote:
> > This introduces a set of bindings for working with iosys_map in
> > rust code
> > using the new Io traits.
> >
> > Signed-off-by: Lyude Paul <lyude@redhat.com>
> >
> > ---
> > V5:
> > - Fix incorrect field size being passed to iosys_map_memcpy_to()
> > - Add an additional unit test, basic_macro(), which can
> > successfully catch
> > the above issue so it doesn't happen again in the future.
> > V6:
> > - Drop as_slice/as_mut_slice (Alice Rhyl)
> > V7:
> > - Start using Alexandre Courbot's wonderful Io, IoCapable and
> > IoKnownSize
> > traits instead of trying to roll our own IO accessors. This also
> > changes
> > the following:
> > - We don't have a custom AsBytes/FromBytes type that we carry
> > around any
> > longer with maps
> > - We now have optional compile-time size checking
> > - We don't need our own unit tests anymore
> > - RawIoSysMap can be used for unsafe IO operations, because why
> > not.
> > - IoSysMapRef::new() can fail now since it needs to ensure SIZE
> > is valid.
> > - We don't implement Deref<RawIoSysMap> for IoSysMapRef any
> > longer. The
> > main reason for this is that we want to avoid users having to
> > manually
> > specify if they want the RawIoSysMap or IoSysMapRef variant
> > functions
> > like io_read()/io_write().
> > This is fine IMHO, but to make sure things remain easy for APIs
> > that
> > wrap around iosys map we make IoSysMapRef.raw_map pub(crate).
> >
> > rust/helpers/helpers.c | 1 +
> > rust/helpers/iosys_map.c | 34 +++++++
> > rust/kernel/iosys_map.rs | 211
> > +++++++++++++++++++++++++++++++++++++++
> > rust/kernel/lib.rs | 1 +
> > 4 files changed, 247 insertions(+)
> > create mode 100644 rust/helpers/iosys_map.c
> > create mode 100644 rust/kernel/iosys_map.rs
> >
> > diff --git a/rust/helpers/helpers.c b/rust/helpers/helpers.c
> > index 1d3333cc0d2a8..bd8ad237aa02e 100644
> > --- a/rust/helpers/helpers.c
> > +++ b/rust/helpers/helpers.c
> > @@ -31,6 +31,7 @@
> > #include "irq.c"
> > #include "fs.c"
> > #include "io.c"
> > +#include "iosys_map.c"
> > #include "jump_label.c"
> > #include "kunit.c"
> > #include "maple_tree.c"
> > diff --git a/rust/helpers/iosys_map.c b/rust/helpers/iosys_map.c
> > new file mode 100644
> > index 0000000000000..6861d4ec48a4b
> > --- /dev/null
> > +++ b/rust/helpers/iosys_map.c
> > @@ -0,0 +1,34 @@
> > +// SPDX-License-Identifier: GPL-2.0
> > +
> > +#include <linux/iosys-map.h>
> > +#include <linux/types.h>
> > +
> > +#define
> > rust_iosys_map_rd(type__)
> > \
> > + __rust_helper
> > type__ \
> > + rust_helper_iosys_map_rd_ ## type__(const struct iosys_map
> > *map, size_t offset) \
> > + {
> > \
> > + return iosys_map_rd(map, offset,
> > type__); \
> > + }
> > +#define
> > rust_iosys_map_wr(type__)
> > \
> > + __rust_helper
> > void \
> > + rust_helper_iosys_map_wr_ ## type__(const struct iosys_map
> > *map, size_t offset, \
> > + type__
> > value) \
> > + {
> > \
> > + iosys_map_wr(map, offset, type__,
> > value); \
> > + }
> > +
> > +rust_iosys_map_rd(u8);
> > +rust_iosys_map_rd(u16);
> > +rust_iosys_map_rd(u32);
> > +
> > +rust_iosys_map_wr(u8);
> > +rust_iosys_map_wr(u16);
> > +rust_iosys_map_wr(u32);
> > +
> > +#ifdef CONFIG_64BIT
> > +rust_iosys_map_rd(u64);
> > +rust_iosys_map_wr(u64);
> > +#endif
> > +
> > +#undef rust_iosys_map_rd
> > +#undef rust_iosys_map_wr
> > diff --git a/rust/kernel/iosys_map.rs b/rust/kernel/iosys_map.rs
> > new file mode 100644
> > index 0000000000000..2070f0fb42cb8
> > --- /dev/null
> > +++ b/rust/kernel/iosys_map.rs
> > @@ -0,0 +1,211 @@
> > +// SPDX-License-Identifier: GPL-2.0
> > +
> > +//! IO-agnostic memory mapping interfaces.
> > +//!
> > +//! This crate provides bindings for the `struct iosys_map` type,
> > which provides a common interface
> > +//! for memory mappings which can reside within coherent memory,
> > or within IO memory.
> > +//!
> > +//! C header: [`include/linux/iosys-
> > map.h`](srctree/include/linux/pci.h)
> > +
> > +use crate::{
> > + error::code::*,
> > + io::{
> > + Io,
> > + IoCapable,
> > + IoKnownSize, //
> > + },
> > + prelude::*, //
> > +};
> > +use bindings;
> > +use core::marker::PhantomData;
> > +
> > +/// Raw unsized representation of a `struct iosys_map`.
> > +///
> > +/// This struct is a transparent wrapper around `struct
> > iosys_map`. The C API does not provide the
> > +/// size of the mapping by default, and thus this type also does
> > not include the size of the
> > +/// mapping. As such, it cannot be used for actually accessing the
> > underlying data pointed to by the
> > +/// mapping.
> > +///
> > +/// With the exception of kernel crates which may provide their
> > own wrappers around `RawIoSysMap`,
> > +/// users will typically not interact with this type directly.
> > +#[repr(transparent)]
> > +pub struct RawIoSysMap<const SIZE: usize =
> > 0>(bindings::iosys_map);
> > +
> > +impl<const SIZE: usize> RawIoSysMap<SIZE> {
> > + /// Convert from a raw `bindings::iosys_map`.
> > + #[expect(unused)]
> > + #[inline]
> > + pub(crate) fn from_raw(val: bindings::iosys_map) -> Self {
> > + Self(val)
> > + }
> > +
> > + /// Returns whether the mapping is within IO memory space or
> > not.
> > + #[inline]
> > + pub fn is_iomem(&self) -> bool {
> > + self.0.is_iomem
> > + }
> > +
> > + /// Convert from a `RawIoSysMap<SIZE>` to a raw
> > `bindings::iosys_map` ref.
> > + #[expect(unused)]
> > + #[inline]
> > + pub(crate) fn as_raw(&self) -> &bindings::iosys_map {
> > + &self.0
> > + }
> > +
> > + /// Convert from a `RawIoSysMap<SIZE>` to a raw mutable
> > `bindings::iosys_map` ref.
> > + #[allow(unused)]
> > + #[inline]
> > + pub(crate) fn as_raw_mut(&mut self) -> &mut
> > bindings::iosys_map {
> > + &mut self.0
> > + }
> > +
> > + /// Returns the address pointed to by this iosys map.
> > + ///
> > + /// Note that this address is not guaranteed to be valid, and
> > may or may not reside in I/O
> > + /// memory.
> > + #[inline]
> > + pub fn addr(&self) -> usize {
> > + (if self.is_iomem() {
> > + // SAFETY: We confirmed above that this iosys map is
> > contained within iomem, so it's
> > + // safe to read vaddr_iomem
> > + unsafe { self.0.__bindgen_anon_1.vaddr_iomem }
> > + } else {
> > + // SAFETY: We confirmed above that this iosys map is
> > not contaned within iomem, so it's
> > + // safe to read vaddr.
> > + unsafe { self.0.__bindgen_anon_1.vaddr }
> > + }) as usize
> > + }
> > +}
> > +
> > +// SAFETY: As we make no guarantees about the validity of the
> > mapping, there's no issue with sending
> > +// this type between threads.
> > +unsafe impl<const SIZE: usize> Send for RawIoSysMap<SIZE> {}
> > +
> > +impl<const SIZE: usize> Clone for RawIoSysMap<SIZE> {
> > + fn clone(&self) -> Self {
> > + Self(self.0)
> > + }
> > +}
> > +
> > +macro_rules! impl_raw_iosys_map_io_capable {
> > + ($ty:ty, $read_fn:ident, $write_fn:ident) => {
> > + impl<const SIZE: usize> IoCapable<$ty> for
> > RawIoSysMap<SIZE> {
> > + #[inline(always)]
> > + unsafe fn io_read(&self, address: usize) -> $ty {
> > + // SAFETY: By the trait invariant `address` is a
> > valid address for iosys map
> > + // operations.
> > + unsafe { bindings::$read_fn(&self.0, address) }
> > + }
> > +
> > + #[inline(always)]
> > + unsafe fn io_write(&self, value: $ty, address: usize)
> > {
> > + // SAFETY: By the trait invariant `address` is a
> > valid address for iosys map
> > + // operations.
> > + unsafe { bindings::$write_fn(&self.0, address,
> > value) };
> > + }
> > + }
> > + };
> > +}
> > +
>
> I think there might be a mismatch between the absolute address being
> passed to these read and write functions and the bindings helpers
> that expect an offset argument.
>
> This crashed Tyr when I tried to write the firmware in u8 chunks at
> incremental offsets. But if i just calculated the offset and passed
> that
> instead of the absolute address, this worked fine:
>
> let offset = address - self.addr();
> unsafe { bindings::$write_fn(&self.0, offset, value) };
>
> Here's some of the call trace:
>
> [ 31.553727] tyr fb000000.gpu: supply sram not found, using dummy
> regulator
> [ 31.555096] tyr fb000000.gpu: mali-unknown id 0xa867 major 0x67
> minor 0x0 status 0x5
> [ 31.555778] tyr fb000000.gpu: Features: L2:0x7120306 Tiler:0x809
> Mem:0x301 MMU:0x2830 AS:0xff
> [ 31.556521] tyr fb000000.gpu: shader_present=0x0000000000050005
> l2_present=0x0000000000000001 tiler_present=0x0000000000000001
> [ 31.557868] stackdepot: allocating hash table of 524288 entries
> via kvcalloc
> [ OK ] Started systemd-tmpfiles-clean.tim…y Cleanup of Temporary
> Directories.
> [ 31.562424] stackdepot: allocating space for 8192 stack pools via
> kvcalloc
> [ OK ] Reached target timers.target - Timer Units.
> [ 31.563676] tyr: Firmware protected mode entry not supported,
> ignoring
> [ 31.571391] Unable to handle kernel paging request at virtual
> address 0000000000800069
> [ 31.572762] Mem abort info:
> [ 31.573027] ESR = 0x0000000096000021
> [ 31.573402] EC = 0x25: DABT (current EL), IL = 32 bits
> [ 31.573878] SET = 0, FnV = 0
> [ 31.574157] EA = 0, S1PTW = 0
> [ 31.574442] FSC = 0x21: alignment fault
> [ OK ] Listening on dbus.socket - D-Bus System Message Bus Socket.
> [ 31.593348] Data abort info:
> [ 31.593628] ISV = 0, ISS = 0x00000021, ISS2 = 0x00000000
> [ 31.594117] CM = 0, WnR = 0, TnD = 0, TagAccess = 0
> [ 31.594567] GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0
> [ 31.595042] user pgtable: 4k pages, 48-bit VAs,
> pgdp=0000000105e50000
> [ 31.595613] [0000000000800069] pgd=0000000000000000,
> p4d=0000000000000000
> [ 31.596434] Internal error: Oops: 0000000096000021 [#1] SMP
> [ 31.596936] Modules linked in: snd tyr(+) soundcore sha256
> cfg80211 rfkill pci_endpoint_test fuse dm_mod ipv6
> [ 31.597830] CPU: 3 UID: 0 PID: 241 Comm: (udev-worker) Not tainted
> 6.19.0-rc5 #282 PREEMPT
> [ 31.598561] Hardware name: Radxa ROCK 5B (DT)
> [ 31.598944] pstate: 00400009 (nzcv daif +PAN -UAO -TCO -DIT -SSBS
> BTYPE=--)
> [ 31.599554] pc : __d_lookup_rcu+0xbc/0x168
> [ 31.599921] lr : __d_lookup_rcu+0x60/0x168
> [ 31.600283] sp : ffff800081ebba00
> [ 31.600574] x29: ffff800081ebba00 x28: 0000000000000049 x27:
> ffff00010d07002b
> [ 31.601205] x26: 0000000000000081 x25: 000000000006915e x24:
> ffff0002f6600000
> [ 31.601848] x23: ffff00010396a040 x22: ffff000102098460 x21:
> 000000036915e207
> [ 31.601859] x20: ffff000101e10af8 x19: ffff800081ebbcac x18:
> 2f64662f666c6573
> [ 31.601867] x17: ffffff00666c6573 x16: 00000000fffffffc x15:
> 0000000000003431
> [ 31.601875] x14: 000000000000000f x13: 0080205100800107 x12:
> 0000000000800069
> [ 31.601882] x11: ffffffffffffffff x10: 0000000000000018 x9 :
> 0000000000000003
> [ 31.601890] x8 : 00000000008000a1 x7 : ffffb9d13df173c8 x6 :
> 0000000000000000
> [ 31.601897] x5 : 0000000000000000 x4 : 0000000000000010 x3 :
> ffffffffffff0a00
> [ 31.601905] x2 : ffff800081ebbcac x1 : ffffb9d1406be718 x0 :
> 0000000000000001
> [ 31.601913] Call trace:
> [ 31.601915] __d_lookup_rcu+0xbc/0x168 (P)
> [ 31.601922] lookup_fast+0x98/0x174
> [ 31.601929] link_path_walk+0x280/0x528
> [ 31.601935] path_lookupat+0x60/0x1f0
> [ 31.601941] do_o_path+0x34/0xb4
> [ 31.601947] path_openat+0xccc/0xe34
> [ 31.601953] do_filp_open+0xc0/0x170
> [ 31.601958] do_sys_openat2+0x88/0x10c
> [ 31.601963] __arm64_sys_openat+0x70/0x9c
> [ 31.601968] invoke_syscall+0x40/0xcc
> [ 31.601974] el0_svc_common+0x80/0xd8
> [ 31.601979] do_el0_svc+0x1c/0x28
> [ 31.601984] el0_svc+0x54/0x1d4
> [ 31.601991] el0t_64_sync_handler+0x84/0x12c
> [ 31.601997] el0t_64_sync+0x198/0x19c
> [ 31.602005] Code: 14000003 f9400108 b4000428 d100e10c (88dffd8c)
> [ 31.602009] ---[ end trace 0000000000000000 ]---
> [ OK ] Listening on sshd-unix-local.socke…temd-ssh-generator,
> AF_UNIX Local).
> [ 31.620067] mc: Linux media interface: v0.10
> [ 31.623483] [drm] Initialized panthor 1.5.0 for fb000000.gpu on
> minor 0
> [ 31.623765] Unable to handle kernel paging request at virtual
> address 00802a4d0080284d
> [ 31.624285] tyr fb000000.gpu: Tyr initialized correctly.
> [ 31.624752] Mem abort info:
> [ 31.624754] ESR = 0x0000000096000004
> [ 31.625847] EC = 0x25: DABT (current EL), IL = 32 bits
> [ 31.626310] SET = 0, FnV = 0
> [ 31.626578] EA = 0, S1PTW = 0
> [ 31.626853] FSC = 0x04: level 0 translation fault
> [ 31.627277] Data abort info:
> [ 31.627529] ISV = 0, ISS = 0x00000004, ISS2 = 0x00000000
> [ 31.628006] CM = 0, WnR = 0, TnD = 0, TagAccess = 0
> [ 31.628447] GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0
> [ 31.628910] [00802a4d0080284d] address between user and kernel
> address ranges
> [ 31.629535] Internal error: Oops: 0000000096000004 [#2] SMP
> [ 31.630030] Modules linked in: mc drm_client_lib snd tyr soundcore
> sha256 cfg80211 rfkill pci_endpoint_test fuse dm_mod ipv6
> [ 31.631017] CPU: 4 UID: 0 PID: 225 Comm: systemd-udevd Tainted:
> G D 6.19.0-rc5 #282 PREEMPT
> [ 31.631877] Tainted: [D]=DIE
> [ 31.632129] Hardware name: Radxa ROCK 5B (DT)
> [ 31.632506] pstate: a0400009 (NzCv daif +PAN -UAO -TCO -DIT -SSBS
> BTYPE=--)
> [ 31.633111] pc : ___d_drop+0xd8/0x144
> [ 31.633433] lr : d_invalidate+0x3c/0x110
> [ 31.633776] sp : ffff800081d3ba70
> [ 31.634064] x29: ffff800081d3ba90 x28: 0000000000000001 x27:
> ffffb9d140bfa000
> [ 31.634685] x26: ffff0001039b5108 x25: ffff00010396a000 x24:
> ffffb9d140166275
> [ 31.635305] x23: ffffb9d140bfa000 x22: ffffb9d140152f3a x21:
> ffff000104076000
> [ 31.635925] x20: ffff00010e520098 x19: ffff00010396a000 x18:
> 0000000000000000
> [ 31.636545] x17: 0000000000000000 x16: 0000000000000000 x15:
> 0000000000000000
> [ 31.637165] x14: 0000000000000000 x13: ffff00010792b8f0 x12:
> 000000000000017a
> [ 31.637785] x11: 00000000008000a1 x10: 00802a4d0080284d x9 :
> ffff0002f6604010
> [ 31.638405] x8 : ffff000105bfcd40 x7 : 0000000000000000 x6 :
> ffffb9d13e272c60
> [ 31.639026] x5 : 0000000000000000 x4 : 0000000000000001 x3 :
> 0000000000000000
> [ 31.639646] x2 : 0000000000000000 x1 : 0000000000000000 x0 :
> ffff00010396a000
> [ 31.640266] Call trace:
> [ 31.640479] ___d_drop+0xd8/0x144 (P)
> [ 31.640800] d_invalidate+0x3c/0x110
> [ 31.641113] proc_invalidate_siblings_dcache+0x244/0x2b8
> [ 31.641577] proc_flush_pid+0x1c/0x28
> [ 31.641897] release_task+0x560/0x668
> [ 31.642218] wait_consider_task+0x5a0/0xb44
> [ 31.642582] __do_wait+0x154/0x1f0
> [ 31.642879] do_wait+0x84/0x16c
> [ 31.643154] __arm64_sys_waitid+0xac/0x218
> [ 31.643512] invoke_syscall+0x40/0xcc
> [ 31.643833] el0_svc_common+0x80/0xd8
> [ 31.644152] do_el0_svc+0x1c/0x28
> [ 31.644441] el0_svc+0x54/0x1d4
> [ 31.644718] el0t_64_sync_handler+0x84/0x12c
> [ 31.645090] el0t_64_sync+0x198/0x19c
> [ 31.645412] Code: 5280002a f85f83a0 17fffff1 a944280b (f940014c)
> [ 31.645941] ---[ end trace 0000000000000000 ]---
>
> thanks,
> Deborah
^ permalink raw reply [flat|nested] 16+ messages in thread* Re: [PATCH v7 6/7] rust: Introduce iosys_map bindings
2026-03-04 22:59 ` lyude
@ 2026-03-06 20:59 ` Deborah Brouwer
0 siblings, 0 replies; 16+ messages in thread
From: Deborah Brouwer @ 2026-03-06 20:59 UTC (permalink / raw)
To: lyude
Cc: linux-kernel, dri-devel, rust-for-linux, Danilo Krummrich,
nouveau, Daniel Almeida, Gary Guo, Benno Lossin,
Alexandre Courbot, Janne Grunau
On Wed, Mar 04, 2026 at 05:59:18PM -0500, lyude@redhat.com wrote:
> Hey! Sorry for the long delay in response
>
> Would you mind sending me the size/offset numbers you used here? My
> unit tests should have caught issues like this, so if they're happening
> I definitely should add it to the testcases I've got to make sure
> there's no problems
Sure, here's some more details of the firmware sections I am loading:
user@trixie:~$ dmesg | grep -i tyr
[ 34.380056] tyr fb000000.gpu: supply sram not found, using dummy regulator
[ 34.392969] tyr fb000000.gpu: mali-g610 id 0xa867 major 0x0 minor 0x0 status 0x5
[ 34.393625] tyr fb000000.gpu: Features: L2:0x7120306 Tiler:0x809 Mem:0x301 MMU:0x2830 AS:0xff
[ 34.394408] tyr fb000000.gpu: shader_present=0x0000000000050005 l2_present=0x0000000000000001 tiler_present=0x0000000000000001
[ 34.400801] tyr: Firmware protected mode entry not supported, ignoring
[ 34.401265] tyr: fw: Section 0
[ 34.403324] tyr: fw: init_section_mem()
[ 34.403331] tyr: Base address of vmap: 0xffff800081c7f000
[ 34.403936] tyr: Size of allocated GEM buffer object (mem.bo.size()): 0x1000
[ 34.405314] tyr: Firmware data length: 0x7c
[ 34.406627] tyr: fw: try_write8 - offset: 0x0, byte: 0x0
[ 34.407262] tyr: fw: try_write8 - offset: 0x1, byte: 0x0
[ 34.407901] tyr: fw: try_write8 - offset: 0x2, byte: 0x3
[ 34.408996] tyr: fw: try_write8 - offset: 0x3, byte: 0x2
[ 34.410050] tyr: fw: try_write8 - offset: 0x4, byte: 0x8d
[ 34.410858] tyr: fw: try_write8 - offset: 0x5, byte: 0x1f
[ 34.411940] tyr: fw: try_write8 - offset: 0x6, byte: 0x80
[ 34.412857] tyr: fw: try_write8 - offset: 0x7, byte: 0x0
[ 34.413747] tyr: fw: try_write8 - offset: 0x8, byte: 0x51
[ 34.413752] tyr: fw: try_write8 - offset: 0x9, byte: 0x20
[ 34.413924] tyr: fw: Section 1
[ 34.413940] tyr: fw: init_section_mem()
[ 34.413943] tyr: Base address of vmap: 0xffff800081d1f000
[ 34.413947] tyr: Size of allocated GEM buffer object (mem.bo.size()): 0x2000
[ 34.413950] tyr: Firmware data length: 0x12dc
[ 34.413952] tyr: fw: try_write8 - offset: 0x0, byte: 0x0
[ 34.413955] tyr: fw: try_write8 - offset: 0x1, byte: 0x0
[ 34.413958] tyr: fw: try_write8 - offset: 0x2, byte: 0x0
[ 34.413961] tyr: fw: try_write8 - offset: 0x3, byte: 0x0
[ 34.413964] tyr: fw: try_write8 - offset: 0x4, byte: 0x0
[ 34.413967] tyr: fw: try_write8 - offset: 0x5, byte: 0x0
[ 34.413969] tyr: fw: try_write8 - offset: 0x6, byte: 0x0
[ 34.413972] tyr: fw: try_write8 - offset: 0x7, byte: 0x0
[ 34.413975] tyr: fw: try_write8 - offset: 0x8, byte: 0x0
[ 34.413977] tyr: fw: try_write8 - offset: 0x9, byte: 0x0
[ 34.414043] tyr: fw: Section 2
[ 34.414050] tyr: fw: init_section_mem()
[ 34.414052] tyr: Base address of vmap: 0xffff800081d13000
[ 34.414055] tyr: Size of allocated GEM buffer object (mem.bo.size()): 0x1000
[ 34.414058] tyr: Firmware data length: 0x7c
[ 34.414061] tyr: fw: try_write8 - offset: 0x0, byte: 0x0
[ 34.414064] tyr: fw: try_write8 - offset: 0x1, byte: 0x0
[ 34.414066] tyr: fw: try_write8 - offset: 0x2, byte: 0x3
[ 34.414069] tyr: fw: try_write8 - offset: 0x3, byte: 0x2
[ 34.414072] tyr: fw: try_write8 - offset: 0x4, byte: 0x8d
[ 34.414074] tyr: fw: try_write8 - offset: 0x5, byte: 0x1f
[ 34.414077] tyr: fw: try_write8 - offset: 0x6, byte: 0x80
[ 34.414080] tyr: fw: try_write8 - offset: 0x7, byte: 0x0
[ 34.414082] tyr: fw: try_write8 - offset: 0x8, byte: 0x51
[ 34.414085] tyr: fw: try_write8 - offset: 0x9, byte: 0x20
[ 34.414346] tyr: fw: Section 3
[ 34.414358] tyr: fw: init_section_mem()
[ 34.414361] tyr: Base address of vmap: 0xffff800083330000
[ 34.414364] tyr: Size of allocated GEM buffer object (mem.bo.size()): 0x20000
[ 34.414367] tyr: Firmware data length: 0x12ff0
[ 34.414370] tyr: fw: try_write8 - offset: 0x0, byte: 0xdf
[ 34.414372] tyr: fw: try_write8 - offset: 0x1, byte: 0xf8
[ 34.414375] tyr: fw: try_write8 - offset: 0x2, byte: 0xc
[ 34.414378] tyr: fw: try_write8 - offset: 0x3, byte: 0xd0
[ 34.414381] tyr: fw: try_write8 - offset: 0x4, byte: 0x2
[ 34.414384] tyr: fw: try_write8 - offset: 0x5, byte: 0xf0
[ 34.414386] tyr: fw: try_write8 - offset: 0x6, byte: 0x82
[ 34.414389] tyr: fw: try_write8 - offset: 0x7, byte: 0xfb
[ 34.414392] tyr: fw: try_write8 - offset: 0x8, byte: 0x0
[ 34.414395] tyr: fw: try_write8 - offset: 0x9, byte: 0x48
[ 34.415676] tyr: fw: Section 4
[ 34.446336] tyr: fw: init_section_mem()
[ 34.446345] tyr: Base address of vmap: 0xffff8000833a0000
[ 34.447001] tyr: Size of allocated GEM buffer object (mem.bo.size()): 0x40000
[ 34.447907] tyr: Firmware data length: 0x2b118
[ 34.449505] tyr: fw: try_write8 - offset: 0x0, byte: 0x0
[ 34.450342] tyr: fw: try_write8 - offset: 0x1, byte: 0x0
[ 34.451308] tyr: fw: try_write8 - offset: 0x2, byte: 0x0
[ 34.452244] tyr: fw: try_write8 - offset: 0x3, byte: 0x0
[ 34.453177] tyr: fw: try_write8 - offset: 0x4, byte: 0x0
[ 34.454122] tyr: fw: try_write8 - offset: 0x5, byte: 0x30
[ 34.455382] tyr: fw: try_write8 - offset: 0x6, byte: 0x0
[ 34.457468] tyr: fw: try_write8 - offset: 0x7, byte: 0x4
[ 34.458417] tyr: fw: try_write8 - offset: 0x8, byte: 0x0
[ 34.459111] tyr: fw: try_write8 - offset: 0x9, byte: 0x70
[ 34.461295] tyr: fw: Section 5
[ 34.462043] tyr: fw: init_section_mem()
[ 34.462046] tyr: Base address of vmap: 0xffff8000833e1000
[ 34.462873] tyr: Size of allocated GEM buffer object (mem.bo.size()): 0x40000
[ 34.463793] tyr: Firmware data length: 0x2b118
[ 34.465285] tyr: fw: try_write8 - offset: 0x0, byte: 0x0
[ 34.466230] tyr: fw: try_write8 - offset: 0x1, byte: 0x0
[ 34.467201] tyr: fw: try_write8 - offset: 0x2, byte: 0x0
[ 34.468119] tyr: fw: try_write8 - offset: 0x3, byte: 0x0
[ 34.469064] tyr: fw: try_write8 - offset: 0x4, byte: 0x0
[ 34.470006] tyr: fw: try_write8 - offset: 0x5, byte: 0x30
[ 34.471145] tyr: fw: try_write8 - offset: 0x6, byte: 0x0
[ 34.471901] tyr: fw: try_write8 - offset: 0x7, byte: 0x4
[ 34.472846] tyr: fw: try_write8 - offset: 0x8, byte: 0x0
[ 34.473597] tyr: fw: try_write8 - offset: 0x9, byte: 0x70
[ 34.475292] tyr: fw: Section 6
[ 34.475752] tyr: fw: init_section_mem()
[ 34.475756] tyr: Base address of vmap: 0xffff800083422000
[ 34.476570] tyr: Size of allocated GEM buffer object (mem.bo.size()): 0xc000
[ 34.477042] tyr: Firmware data length: 0x2000
[ 34.478676] tyr: fw: try_write8 - offset: 0x0, byte: 0x0
[ 34.479459] tyr: fw: try_write8 - offset: 0x1, byte: 0x0
[ 34.479462] tyr: fw: try_write8 - offset: 0x2, byte: 0x0
[ 34.479465] tyr: fw: try_write8 - offset: 0x3, byte: 0x0
[ 34.479468] tyr: fw: try_write8 - offset: 0x4, byte: 0x0
[ 34.479471] tyr: fw: try_write8 - offset: 0x5, byte: 0x0
[ 34.482247] tyr: fw: try_write8 - offset: 0x6, byte: 0x0
[ 34.482820] tyr: fw: try_write8 - offset: 0x7, byte: 0x0
[ 34.483289] tyr: fw: try_write8 - offset: 0x8, byte: 0x0
[ 34.483291] tyr: fw: try_write8 - offset: 0x9, byte: 0x20
[ 34.507842] tyr fb000000.gpu: Tyr initialized correctly.
Here's a link to the commit where I added these print statements
if you would like to see more context:
https://gitlab.freedesktop.org/dbrouwer/linux/-/commit/2232b848a9ceef681382159c9b1471e6df0c00d6
Let me know if I can help more or test anything.
Thanks,
Deborah
>
> On Thu, 2026-02-12 at 17:59 -0800, Deborah Brouwer wrote:
> > Hi Lyude,
> >
> > On Fri, Feb 06, 2026 at 05:34:30PM -0500, Lyude Paul wrote:
> > > This introduces a set of bindings for working with iosys_map in
> > > rust code
> > > using the new Io traits.
> > >
> > > Signed-off-by: Lyude Paul <lyude@redhat.com>
> > >
> > > ---
> > > V5:
> > > - Fix incorrect field size being passed to iosys_map_memcpy_to()
> > > - Add an additional unit test, basic_macro(), which can
> > > successfully catch
> > > the above issue so it doesn't happen again in the future.
> > > V6:
> > > - Drop as_slice/as_mut_slice (Alice Rhyl)
> > > V7:
> > > - Start using Alexandre Courbot's wonderful Io, IoCapable and
> > > IoKnownSize
> > > traits instead of trying to roll our own IO accessors. This also
> > > changes
> > > the following:
> > > - We don't have a custom AsBytes/FromBytes type that we carry
> > > around any
> > > longer with maps
> > > - We now have optional compile-time size checking
> > > - We don't need our own unit tests anymore
> > > - RawIoSysMap can be used for unsafe IO operations, because why
> > > not.
> > > - IoSysMapRef::new() can fail now since it needs to ensure SIZE
> > > is valid.
> > > - We don't implement Deref<RawIoSysMap> for IoSysMapRef any
> > > longer. The
> > > main reason for this is that we want to avoid users having to
> > > manually
> > > specify if they want the RawIoSysMap or IoSysMapRef variant
> > > functions
> > > like io_read()/io_write().
> > > This is fine IMHO, but to make sure things remain easy for APIs
> > > that
> > > wrap around iosys map we make IoSysMapRef.raw_map pub(crate).
> > >
> > > rust/helpers/helpers.c | 1 +
> > > rust/helpers/iosys_map.c | 34 +++++++
> > > rust/kernel/iosys_map.rs | 211
> > > +++++++++++++++++++++++++++++++++++++++
> > > rust/kernel/lib.rs | 1 +
> > > 4 files changed, 247 insertions(+)
> > > create mode 100644 rust/helpers/iosys_map.c
> > > create mode 100644 rust/kernel/iosys_map.rs
> > >
> > > diff --git a/rust/helpers/helpers.c b/rust/helpers/helpers.c
> > > index 1d3333cc0d2a8..bd8ad237aa02e 100644
> > > --- a/rust/helpers/helpers.c
> > > +++ b/rust/helpers/helpers.c
> > > @@ -31,6 +31,7 @@
> > > #include "irq.c"
> > > #include "fs.c"
> > > #include "io.c"
> > > +#include "iosys_map.c"
> > > #include "jump_label.c"
> > > #include "kunit.c"
> > > #include "maple_tree.c"
> > > diff --git a/rust/helpers/iosys_map.c b/rust/helpers/iosys_map.c
> > > new file mode 100644
> > > index 0000000000000..6861d4ec48a4b
> > > --- /dev/null
> > > +++ b/rust/helpers/iosys_map.c
> > > @@ -0,0 +1,34 @@
> > > +// SPDX-License-Identifier: GPL-2.0
> > > +
> > > +#include <linux/iosys-map.h>
> > > +#include <linux/types.h>
> > > +
> > > +#define
> > > rust_iosys_map_rd(type__)
> > > \
> > > + __rust_helper
> > > type__ \
> > > + rust_helper_iosys_map_rd_ ## type__(const struct iosys_map
> > > *map, size_t offset) \
> > > + {
> > > \
> > > + return iosys_map_rd(map, offset,
> > > type__); \
> > > + }
> > > +#define
> > > rust_iosys_map_wr(type__)
> > > \
> > > + __rust_helper
> > > void \
> > > + rust_helper_iosys_map_wr_ ## type__(const struct iosys_map
> > > *map, size_t offset, \
> > > + type__
> > > value) \
> > > + {
> > > \
> > > + iosys_map_wr(map, offset, type__,
> > > value); \
> > > + }
> > > +
> > > +rust_iosys_map_rd(u8);
> > > +rust_iosys_map_rd(u16);
> > > +rust_iosys_map_rd(u32);
> > > +
> > > +rust_iosys_map_wr(u8);
> > > +rust_iosys_map_wr(u16);
> > > +rust_iosys_map_wr(u32);
> > > +
> > > +#ifdef CONFIG_64BIT
> > > +rust_iosys_map_rd(u64);
> > > +rust_iosys_map_wr(u64);
> > > +#endif
> > > +
> > > +#undef rust_iosys_map_rd
> > > +#undef rust_iosys_map_wr
> > > diff --git a/rust/kernel/iosys_map.rs b/rust/kernel/iosys_map.rs
> > > new file mode 100644
> > > index 0000000000000..2070f0fb42cb8
> > > --- /dev/null
> > > +++ b/rust/kernel/iosys_map.rs
> > > @@ -0,0 +1,211 @@
> > > +// SPDX-License-Identifier: GPL-2.0
> > > +
> > > +//! IO-agnostic memory mapping interfaces.
> > > +//!
> > > +//! This crate provides bindings for the `struct iosys_map` type,
> > > which provides a common interface
> > > +//! for memory mappings which can reside within coherent memory,
> > > or within IO memory.
> > > +//!
> > > +//! C header: [`include/linux/iosys-
> > > map.h`](srctree/include/linux/pci.h)
> > > +
> > > +use crate::{
> > > + error::code::*,
> > > + io::{
> > > + Io,
> > > + IoCapable,
> > > + IoKnownSize, //
> > > + },
> > > + prelude::*, //
> > > +};
> > > +use bindings;
> > > +use core::marker::PhantomData;
> > > +
> > > +/// Raw unsized representation of a `struct iosys_map`.
> > > +///
> > > +/// This struct is a transparent wrapper around `struct
> > > iosys_map`. The C API does not provide the
> > > +/// size of the mapping by default, and thus this type also does
> > > not include the size of the
> > > +/// mapping. As such, it cannot be used for actually accessing the
> > > underlying data pointed to by the
> > > +/// mapping.
> > > +///
> > > +/// With the exception of kernel crates which may provide their
> > > own wrappers around `RawIoSysMap`,
> > > +/// users will typically not interact with this type directly.
> > > +#[repr(transparent)]
> > > +pub struct RawIoSysMap<const SIZE: usize =
> > > 0>(bindings::iosys_map);
> > > +
> > > +impl<const SIZE: usize> RawIoSysMap<SIZE> {
> > > + /// Convert from a raw `bindings::iosys_map`.
> > > + #[expect(unused)]
> > > + #[inline]
> > > + pub(crate) fn from_raw(val: bindings::iosys_map) -> Self {
> > > + Self(val)
> > > + }
> > > +
> > > + /// Returns whether the mapping is within IO memory space or
> > > not.
> > > + #[inline]
> > > + pub fn is_iomem(&self) -> bool {
> > > + self.0.is_iomem
> > > + }
> > > +
> > > + /// Convert from a `RawIoSysMap<SIZE>` to a raw
> > > `bindings::iosys_map` ref.
> > > + #[expect(unused)]
> > > + #[inline]
> > > + pub(crate) fn as_raw(&self) -> &bindings::iosys_map {
> > > + &self.0
> > > + }
> > > +
> > > + /// Convert from a `RawIoSysMap<SIZE>` to a raw mutable
> > > `bindings::iosys_map` ref.
> > > + #[allow(unused)]
> > > + #[inline]
> > > + pub(crate) fn as_raw_mut(&mut self) -> &mut
> > > bindings::iosys_map {
> > > + &mut self.0
> > > + }
> > > +
> > > + /// Returns the address pointed to by this iosys map.
> > > + ///
> > > + /// Note that this address is not guaranteed to be valid, and
> > > may or may not reside in I/O
> > > + /// memory.
> > > + #[inline]
> > > + pub fn addr(&self) -> usize {
> > > + (if self.is_iomem() {
> > > + // SAFETY: We confirmed above that this iosys map is
> > > contained within iomem, so it's
> > > + // safe to read vaddr_iomem
> > > + unsafe { self.0.__bindgen_anon_1.vaddr_iomem }
> > > + } else {
> > > + // SAFETY: We confirmed above that this iosys map is
> > > not contaned within iomem, so it's
> > > + // safe to read vaddr.
> > > + unsafe { self.0.__bindgen_anon_1.vaddr }
> > > + }) as usize
> > > + }
> > > +}
> > > +
> > > +// SAFETY: As we make no guarantees about the validity of the
> > > mapping, there's no issue with sending
> > > +// this type between threads.
> > > +unsafe impl<const SIZE: usize> Send for RawIoSysMap<SIZE> {}
> > > +
> > > +impl<const SIZE: usize> Clone for RawIoSysMap<SIZE> {
> > > + fn clone(&self) -> Self {
> > > + Self(self.0)
> > > + }
> > > +}
> > > +
> > > +macro_rules! impl_raw_iosys_map_io_capable {
> > > + ($ty:ty, $read_fn:ident, $write_fn:ident) => {
> > > + impl<const SIZE: usize> IoCapable<$ty> for
> > > RawIoSysMap<SIZE> {
> > > + #[inline(always)]
> > > + unsafe fn io_read(&self, address: usize) -> $ty {
> > > + // SAFETY: By the trait invariant `address` is a
> > > valid address for iosys map
> > > + // operations.
> > > + unsafe { bindings::$read_fn(&self.0, address) }
> > > + }
> > > +
> > > + #[inline(always)]
> > > + unsafe fn io_write(&self, value: $ty, address: usize)
> > > {
> > > + // SAFETY: By the trait invariant `address` is a
> > > valid address for iosys map
> > > + // operations.
> > > + unsafe { bindings::$write_fn(&self.0, address,
> > > value) };
> > > + }
> > > + }
> > > + };
> > > +}
> > > +
> >
> > I think there might be a mismatch between the absolute address being
> > passed to these read and write functions and the bindings helpers
> > that expect an offset argument.
> >
> > This crashed Tyr when I tried to write the firmware in u8 chunks at
> > incremental offsets. But if i just calculated the offset and passed
> > that
> > instead of the absolute address, this worked fine:
> >
> > let offset = address - self.addr();
> > unsafe { bindings::$write_fn(&self.0, offset, value) };
> >
> > Here's some of the call trace:
> >
> > [ 31.553727] tyr fb000000.gpu: supply sram not found, using dummy
> > regulator
> > [ 31.555096] tyr fb000000.gpu: mali-unknown id 0xa867 major 0x67
> > minor 0x0 status 0x5
> > [ 31.555778] tyr fb000000.gpu: Features: L2:0x7120306 Tiler:0x809
> > Mem:0x301 MMU:0x2830 AS:0xff
> > [ 31.556521] tyr fb000000.gpu: shader_present=0x0000000000050005
> > l2_present=0x0000000000000001 tiler_present=0x0000000000000001
> > [ 31.557868] stackdepot: allocating hash table of 524288 entries
> > via kvcalloc
> > [ OK ] Started systemd-tmpfiles-clean.tim…y Cleanup of Temporary
> > Directories.
> > [ 31.562424] stackdepot: allocating space for 8192 stack pools via
> > kvcalloc
> > [ OK ] Reached target timers.target - Timer Units.
> > [ 31.563676] tyr: Firmware protected mode entry not supported,
> > ignoring
> > [ 31.571391] Unable to handle kernel paging request at virtual
> > address 0000000000800069
> > [ 31.572762] Mem abort info:
> > [ 31.573027] ESR = 0x0000000096000021
> > [ 31.573402] EC = 0x25: DABT (current EL), IL = 32 bits
> > [ 31.573878] SET = 0, FnV = 0
> > [ 31.574157] EA = 0, S1PTW = 0
> > [ 31.574442] FSC = 0x21: alignment fault
> > [ OK ] Listening on dbus.socket - D-Bus System Message Bus Socket.
> > [ 31.593348] Data abort info:
> > [ 31.593628] ISV = 0, ISS = 0x00000021, ISS2 = 0x00000000
> > [ 31.594117] CM = 0, WnR = 0, TnD = 0, TagAccess = 0
> > [ 31.594567] GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0
> > [ 31.595042] user pgtable: 4k pages, 48-bit VAs,
> > pgdp=0000000105e50000
> > [ 31.595613] [0000000000800069] pgd=0000000000000000,
> > p4d=0000000000000000
> > [ 31.596434] Internal error: Oops: 0000000096000021 [#1] SMP
> > [ 31.596936] Modules linked in: snd tyr(+) soundcore sha256
> > cfg80211 rfkill pci_endpoint_test fuse dm_mod ipv6
> > [ 31.597830] CPU: 3 UID: 0 PID: 241 Comm: (udev-worker) Not tainted
> > 6.19.0-rc5 #282 PREEMPT
> > [ 31.598561] Hardware name: Radxa ROCK 5B (DT)
> > [ 31.598944] pstate: 00400009 (nzcv daif +PAN -UAO -TCO -DIT -SSBS
> > BTYPE=--)
> > [ 31.599554] pc : __d_lookup_rcu+0xbc/0x168
> > [ 31.599921] lr : __d_lookup_rcu+0x60/0x168
> > [ 31.600283] sp : ffff800081ebba00
> > [ 31.600574] x29: ffff800081ebba00 x28: 0000000000000049 x27:
> > ffff00010d07002b
> > [ 31.601205] x26: 0000000000000081 x25: 000000000006915e x24:
> > ffff0002f6600000
> > [ 31.601848] x23: ffff00010396a040 x22: ffff000102098460 x21:
> > 000000036915e207
> > [ 31.601859] x20: ffff000101e10af8 x19: ffff800081ebbcac x18:
> > 2f64662f666c6573
> > [ 31.601867] x17: ffffff00666c6573 x16: 00000000fffffffc x15:
> > 0000000000003431
> > [ 31.601875] x14: 000000000000000f x13: 0080205100800107 x12:
> > 0000000000800069
> > [ 31.601882] x11: ffffffffffffffff x10: 0000000000000018 x9 :
> > 0000000000000003
> > [ 31.601890] x8 : 00000000008000a1 x7 : ffffb9d13df173c8 x6 :
> > 0000000000000000
> > [ 31.601897] x5 : 0000000000000000 x4 : 0000000000000010 x3 :
> > ffffffffffff0a00
> > [ 31.601905] x2 : ffff800081ebbcac x1 : ffffb9d1406be718 x0 :
> > 0000000000000001
> > [ 31.601913] Call trace:
> > [ 31.601915] __d_lookup_rcu+0xbc/0x168 (P)
> > [ 31.601922] lookup_fast+0x98/0x174
> > [ 31.601929] link_path_walk+0x280/0x528
> > [ 31.601935] path_lookupat+0x60/0x1f0
> > [ 31.601941] do_o_path+0x34/0xb4
> > [ 31.601947] path_openat+0xccc/0xe34
> > [ 31.601953] do_filp_open+0xc0/0x170
> > [ 31.601958] do_sys_openat2+0x88/0x10c
> > [ 31.601963] __arm64_sys_openat+0x70/0x9c
> > [ 31.601968] invoke_syscall+0x40/0xcc
> > [ 31.601974] el0_svc_common+0x80/0xd8
> > [ 31.601979] do_el0_svc+0x1c/0x28
> > [ 31.601984] el0_svc+0x54/0x1d4
> > [ 31.601991] el0t_64_sync_handler+0x84/0x12c
> > [ 31.601997] el0t_64_sync+0x198/0x19c
> > [ 31.602005] Code: 14000003 f9400108 b4000428 d100e10c (88dffd8c)
> > [ 31.602009] ---[ end trace 0000000000000000 ]---
> > [ OK ] Listening on sshd-unix-local.socke…temd-ssh-generator,
> > AF_UNIX Local).
> > [ 31.620067] mc: Linux media interface: v0.10
> > [ 31.623483] [drm] Initialized panthor 1.5.0 for fb000000.gpu on
> > minor 0
> > [ 31.623765] Unable to handle kernel paging request at virtual
> > address 00802a4d0080284d
> > [ 31.624285] tyr fb000000.gpu: Tyr initialized correctly.
> > [ 31.624752] Mem abort info:
> > [ 31.624754] ESR = 0x0000000096000004
> > [ 31.625847] EC = 0x25: DABT (current EL), IL = 32 bits
> > [ 31.626310] SET = 0, FnV = 0
> > [ 31.626578] EA = 0, S1PTW = 0
> > [ 31.626853] FSC = 0x04: level 0 translation fault
> > [ 31.627277] Data abort info:
> > [ 31.627529] ISV = 0, ISS = 0x00000004, ISS2 = 0x00000000
> > [ 31.628006] CM = 0, WnR = 0, TnD = 0, TagAccess = 0
> > [ 31.628447] GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0
> > [ 31.628910] [00802a4d0080284d] address between user and kernel
> > address ranges
> > [ 31.629535] Internal error: Oops: 0000000096000004 [#2] SMP
> > [ 31.630030] Modules linked in: mc drm_client_lib snd tyr soundcore
> > sha256 cfg80211 rfkill pci_endpoint_test fuse dm_mod ipv6
> > [ 31.631017] CPU: 4 UID: 0 PID: 225 Comm: systemd-udevd Tainted:
> > G D 6.19.0-rc5 #282 PREEMPT
> > [ 31.631877] Tainted: [D]=DIE
> > [ 31.632129] Hardware name: Radxa ROCK 5B (DT)
> > [ 31.632506] pstate: a0400009 (NzCv daif +PAN -UAO -TCO -DIT -SSBS
> > BTYPE=--)
> > [ 31.633111] pc : ___d_drop+0xd8/0x144
> > [ 31.633433] lr : d_invalidate+0x3c/0x110
> > [ 31.633776] sp : ffff800081d3ba70
> > [ 31.634064] x29: ffff800081d3ba90 x28: 0000000000000001 x27:
> > ffffb9d140bfa000
> > [ 31.634685] x26: ffff0001039b5108 x25: ffff00010396a000 x24:
> > ffffb9d140166275
> > [ 31.635305] x23: ffffb9d140bfa000 x22: ffffb9d140152f3a x21:
> > ffff000104076000
> > [ 31.635925] x20: ffff00010e520098 x19: ffff00010396a000 x18:
> > 0000000000000000
> > [ 31.636545] x17: 0000000000000000 x16: 0000000000000000 x15:
> > 0000000000000000
> > [ 31.637165] x14: 0000000000000000 x13: ffff00010792b8f0 x12:
> > 000000000000017a
> > [ 31.637785] x11: 00000000008000a1 x10: 00802a4d0080284d x9 :
> > ffff0002f6604010
> > [ 31.638405] x8 : ffff000105bfcd40 x7 : 0000000000000000 x6 :
> > ffffb9d13e272c60
> > [ 31.639026] x5 : 0000000000000000 x4 : 0000000000000001 x3 :
> > 0000000000000000
> > [ 31.639646] x2 : 0000000000000000 x1 : 0000000000000000 x0 :
> > ffff00010396a000
> > [ 31.640266] Call trace:
> > [ 31.640479] ___d_drop+0xd8/0x144 (P)
> > [ 31.640800] d_invalidate+0x3c/0x110
> > [ 31.641113] proc_invalidate_siblings_dcache+0x244/0x2b8
> > [ 31.641577] proc_flush_pid+0x1c/0x28
> > [ 31.641897] release_task+0x560/0x668
> > [ 31.642218] wait_consider_task+0x5a0/0xb44
> > [ 31.642582] __do_wait+0x154/0x1f0
> > [ 31.642879] do_wait+0x84/0x16c
> > [ 31.643154] __arm64_sys_waitid+0xac/0x218
> > [ 31.643512] invoke_syscall+0x40/0xcc
> > [ 31.643833] el0_svc_common+0x80/0xd8
> > [ 31.644152] do_el0_svc+0x1c/0x28
> > [ 31.644441] el0_svc+0x54/0x1d4
> > [ 31.644718] el0t_64_sync_handler+0x84/0x12c
> > [ 31.645090] el0t_64_sync+0x198/0x19c
> > [ 31.645412] Code: 5280002a f85f83a0 17fffff1 a944280b (f940014c)
> > [ 31.645941] ---[ end trace 0000000000000000 ]---
> >
> > thanks,
> > Deborah
>
^ permalink raw reply [flat|nested] 16+ messages in thread
* [PATCH v7 7/7] rust: drm/gem: Add vmap functions to shmem bindings
2026-02-06 22:34 [PATCH v7 0/7] Rust bindings for gem shmem + iosys_map Lyude Paul
` (5 preceding siblings ...)
2026-02-06 22:34 ` [PATCH v7 6/7] rust: Introduce iosys_map bindings Lyude Paul
@ 2026-02-06 22:34 ` Lyude Paul
6 siblings, 0 replies; 16+ messages in thread
From: Lyude Paul @ 2026-02-06 22:34 UTC (permalink / raw)
To: linux-kernel, dri-devel, rust-for-linux, Danilo Krummrich
Cc: nouveau, Daniel Almeida, Gary Guo, Benno Lossin,
Alexandre Courbot, Janne Grunau
One of the more obvious use cases for gem shmem objects is the ability to
create mappings into their contents, specifically iosys mappings. Now that
we've added iosys_map rust bindings to the kernel, let's hook these up in
gem shmem.
Similar to how we handle SGTables, we make sure there's two different types
of mappings: owned mappings (kernel::drm::gem::shmem::VMap) and borrowed
mappings (kernel::drm::gem::shmem::VMapRef).
One last note: we change the #[expect(unused)] for RawIoSysMap::from_raw()
to an #[allow(unused)]. Normally we would simply remove the lint assertion,
however - since shmem is conditionally built, we need allow to avoid
hitting warnings in certain kernel configurations.
Signed-off-by: Lyude Paul <lyude@redhat.com>
---
V7:
* Switch over to the new iosys map bindings that use the Io trait
rust/kernel/drm/gem/shmem.rs | 170 ++++++++++++++++++++++++++++++++++-
rust/kernel/iosys_map.rs | 3 +-
2 files changed, 170 insertions(+), 3 deletions(-)
diff --git a/rust/kernel/drm/gem/shmem.rs b/rust/kernel/drm/gem/shmem.rs
index e511a9b6710e0..604fb10325d1e 100644
--- a/rust/kernel/drm/gem/shmem.rs
+++ b/rust/kernel/drm/gem/shmem.rs
@@ -21,6 +21,7 @@
from_err_ptr,
to_result, //
},
+ iosys_map::*,
prelude::*,
scatterlist,
types::{
@@ -29,13 +30,18 @@
}, //
};
use core::{
+ mem::{
+ self,
+ MaybeUninit, //
+ },
ops::{
Deref,
DerefMut, //
},
- ptr::NonNull,
+ ptr::NonNull, //
};
use gem::{
+ BaseObject,
BaseObjectPrivate,
DriverObject,
IntoGEMObject, //
@@ -216,6 +222,72 @@ pub fn owned_sg_table(&self) -> Result<SGTable<T>> {
_owner: self.into(),
})
}
+
+ /// Attempt to create a [`RawIoSysMap`] from the gem object.
+ fn raw_vmap<const SIZE: usize>(&self) -> Result<RawIoSysMap<SIZE>> {
+ let mut map: MaybeUninit<bindings::iosys_map> = MaybeUninit::uninit();
+
+ // SAFETY: drm_gem_shmem_vmap can be called with the DMA reservation lock held
+ to_result(unsafe {
+ // TODO: see top of file
+ bindings::dma_resv_lock(self.raw_dma_resv(), core::ptr::null_mut());
+ let ret = bindings::drm_gem_shmem_vmap_locked(self.as_shmem(), map.as_mut_ptr());
+ bindings::dma_resv_unlock(self.raw_dma_resv());
+ ret
+ })?;
+
+ // SAFETY: if drm_gem_shmem_vmap did not fail, map is initialized now
+ Ok(unsafe { RawIoSysMap::<SIZE>::from_raw(map.assume_init()) })
+ }
+
+ /// Unmap a [`RawIoSysMap`] from the gem object.
+ ///
+ /// # Safety
+ ///
+ /// - The caller promises that `map` came from a prior call to [`Self::raw_vmap`] on this gem
+ /// object.
+ /// - The caller promises that the memory pointed to by `map` will no longer be accesed through
+ /// this instance.
+ unsafe fn raw_vunmap<const SIZE: usize>(&self, map: &mut RawIoSysMap<SIZE>) {
+ let resv = self.raw_dma_resv();
+
+ // SAFETY:
+ // - This function is safe to call with the DMA reservation lock held
+ // - Our `ARef` is proof that the underlying gem object here is initialized and thus safe to
+ // dereference.
+ unsafe {
+ // TODO: see top of file
+ bindings::dma_resv_lock(resv, core::ptr::null_mut());
+ bindings::drm_gem_shmem_vunmap_locked(self.as_shmem(), map.as_raw_mut());
+ bindings::dma_resv_unlock(resv);
+ }
+ }
+
+ /// Creates and returns a virtual kernel memory mapping for this object.
+ pub fn vmap<const SIZE: usize>(&self) -> Result<VMapRef<'_, T, SIZE>> {
+ let map = self.raw_vmap()?;
+
+ Ok(VMapRef {
+ // SAFETY:
+ // - The size of the vmap is the same as the size of the gem
+ // - The vmap will remain alive until this object is dropped.
+ map: unsafe { IoSysMapRef::<SIZE>::new(map, self.size()) }?,
+ owner: self,
+ })
+ }
+
+ /// Creates and returns an owned reference to a virtual kernel memory mapping for this object.
+ pub fn owned_vmap<const SIZE: usize>(&self) -> Result<VMap<T, SIZE>> {
+ // INVARIANT: We verify here that the mapping is at least of SIZE bytes.
+ if self.size() < SIZE {
+ return Err(EINVAL);
+ }
+
+ Ok(VMap {
+ map: self.raw_vmap()?,
+ owner: self.into(),
+ })
+ }
}
impl<T: DriverObject> Deref for Object<T> {
@@ -267,6 +339,102 @@ impl<T: DriverObject> driver::AllocImpl for Object<T> {
};
}
+/// A borrowed reference to a virtual mapping for a shmem-based GEM object in kernel address space.
+pub struct VMapRef<'a, D: DriverObject, const SIZE: usize = 0> {
+ map: IoSysMapRef<'a, SIZE>,
+ owner: &'a Object<D>,
+}
+
+impl<'a, D: DriverObject, const SIZE: usize> Clone for VMapRef<'a, D, SIZE> {
+ fn clone(&self) -> Self {
+ // SAFETY: We have a successful vmap already, so this can't fail
+ unsafe { self.owner.vmap().unwrap_unchecked() }
+ }
+}
+
+impl<'a, D: DriverObject, const SIZE: usize> Deref for VMapRef<'a, D, SIZE> {
+ type Target = IoSysMapRef<'a, SIZE>;
+
+ fn deref(&self) -> &Self::Target {
+ &self.map
+ }
+}
+
+impl<'a, D: DriverObject, const SIZE: usize> DerefMut for VMapRef<'a, D, SIZE> {
+ fn deref_mut(&mut self) -> &mut Self::Target {
+ &mut self.map
+ }
+}
+
+impl<'a, D: DriverObject, const SIZE: usize> Drop for VMapRef<'a, D, SIZE> {
+ fn drop(&mut self) {
+ // SAFETY: Our existence is proof that this map was previously created using self.owner.
+ unsafe { self.owner.raw_vunmap(&mut self.map.raw_map) };
+ }
+}
+
+/// An owned reference to a virtual mapping for a shmem-based GEM object in kernel address space.
+///
+/// # Invariants
+///
+/// - The size of `owner` is >= SIZE.
+/// - The memory pointed to by `map` is at least as large as `T`.
+/// - The memory pointed to by `map` remains valid at least until this object is dropped.
+pub struct VMap<D: DriverObject, const SIZE: usize = 0> {
+ map: RawIoSysMap<SIZE>,
+ owner: ARef<Object<D>>,
+}
+
+impl<D: DriverObject, const SIZE: usize> Clone for VMap<D, SIZE> {
+ fn clone(&self) -> Self {
+ // SAFETY: We have a successful vmap already, so this can't fail
+ unsafe { self.owner.owned_vmap().unwrap_unchecked() }
+ }
+}
+
+impl<'a, D: DriverObject, const SIZE: usize> From<VMapRef<'a, D, SIZE>> for VMap<D, SIZE> {
+ fn from(value: VMapRef<'a, D, SIZE>) -> Self {
+ let this = Self {
+ map: value.map.raw_map.clone(),
+ owner: value.owner.into(),
+ };
+
+ mem::forget(value);
+ this
+ }
+}
+
+impl<D: DriverObject, const SIZE: usize> VMap<D, SIZE> {
+ /// Return a reference to the iosys map for this `VMap`.
+ #[inline]
+ pub fn get(&self) -> IoSysMapRef<'_, SIZE> {
+ // SAFETY:
+ // - The size of the iosys_map is equivalent to the size of the gem object.
+ // - `size` is >= SIZE according to our type invariants, ensuring that we can never pass an
+ // invalid `size` to `IoSysMapRef::new().
+ unsafe {
+ IoSysMapRef::new(self.map.clone(), self.owner.size()).unwrap_unchecked()
+ }
+ }
+
+ /// Borrows a reference to the object that owns this virtual mapping.
+ pub fn owner(&self) -> &Object<D> {
+ &self.owner
+ }
+}
+
+impl<D: DriverObject, const SIZE: usize> Drop for VMap<D, SIZE> {
+ fn drop(&mut self) {
+ // SAFETY: Our existence is proof that this map was previously created using self.owner
+ unsafe { self.owner.raw_vunmap(&mut self.map) };
+ }
+}
+
+/// SAFETY: `iosys_map` objects are safe to send across threads.
+unsafe impl<D: DriverObject, const SIZE: usize> Send for VMap<D, SIZE> {}
+/// SAFETY: `iosys_map` objects are safe to send across threads.
+unsafe impl<D: DriverObject, const SIZE: usize> Sync for VMap<D, SIZE> {}
+
/// An owned reference to a scatter-gather table of DMA address spans for a GEM shmem object.
///
/// This object holds an owned reference to the underlying GEM shmem object, ensuring that the
diff --git a/rust/kernel/iosys_map.rs b/rust/kernel/iosys_map.rs
index 2070f0fb42cb8..b649d2de83093 100644
--- a/rust/kernel/iosys_map.rs
+++ b/rust/kernel/iosys_map.rs
@@ -33,7 +33,7 @@
impl<const SIZE: usize> RawIoSysMap<SIZE> {
/// Convert from a raw `bindings::iosys_map`.
- #[expect(unused)]
+ #[allow(unused)]
#[inline]
pub(crate) fn from_raw(val: bindings::iosys_map) -> Self {
Self(val)
@@ -139,7 +139,6 @@ impl<'a, const SIZE: usize> IoSysMapRef<'a, SIZE> {
///
/// - The caller guarantees that the mapping is of at least `size` bytes.
/// - The caller guarantees that the mapping remains valid for the lifetime of `'a`.
- #[expect(unused)]
pub(crate) unsafe fn new(map: RawIoSysMap<SIZE>, size: usize) -> Result<Self> {
if size < SIZE {
return Err(EINVAL);
--
2.53.0
^ permalink raw reply related [flat|nested] 16+ messages in thread