From: Ivan Tverdokhleb <mbkchelofficialdev@gmail.com>
To: rust-for-linux@vger.kernel.org
Cc: "Danilo Krummrich" <dakr@kernel.org>,
"Alice Ryhl" <aliceryhl@google.com>,
"Matthew Auld" <matthew.auld@intel.com>,
"Arun Pravin" <arunpravin.paneerselvam@amd.com>,
"Joel Fernandes" <jfern@kernel.org>,
"Miguel Ojeda" <ojeda@kernel.org>,
"Boqun Feng" <boqun@kernel.org>, "Gary Guo" <gary@garyguo.net>,
"Björn Roy Baron" <bjorn3_gh@protonmail.com>,
"Benno Lossin" <lossin@kernel.org>,
"Andreas Hindborg" <a.hindborg@kernel.org>,
"Trevor Gross" <tmgross@umich.edu>,
"Daniel Almeida" <daniel.almeida@collabora.com>,
"Tamir Duberstein" <tamird@kernel.org>,
"Alexandre Courbot" <acourbot@nvidia.com>,
"Onur Özkan" <work@onurozkan.dev>,
dri-devel@lists.freedesktop.org
Subject: [PATCH v2] rust: gpu: buddy: store the allocator inside the Mutex
Date: Thu, 19 Nov 2026 13:18:50 +0200 [thread overview]
Message-ID: <20261119111850.33270-1-mbkchelofficialdev@gmail.com> (raw)
GpuBuddyInner keeps the C gpu_buddy structure next to a separate
Mutex<()>, so nothing in the type system ties the data to the lock.
A TODO asked to replace Mutex<()> with Mutex<Opaque<..>> once
Mutex::new() accepts a fallible initializer.
That is now possible, so put the Opaque<gpu_buddy> directly inside the
Mutex. The allocator can then only be reached through the lock guard,
which lets the compiler enforce that the lock is held. This also
removes the now unneeded `inner` field from GpuBuddyGuard.
Update the safety comments that referred to the removed `inner` field.
Signed-off-by: Ivan Tverdokhleb <mbkchelofficialdev@gmail.com>
---
Changes in v2:
- Dropped patch 1/2 ("rust: sync: lock: allow fallible initializers
in Lock::new"): the same change is already in linux-next
(Philipp Stanner's "rust: sync: Accept errors for Lock payload
data"), as Gary pointed out. This patch now depends on it.
- Rebased on next-20261118.
- Fixed the indentation of the new_mutex!() initializer.
- Build-tested on x86_64 (defconfig with CONFIG_RUST=y,
CONFIG_GPU_BUDDY=y and CONFIG_NOVA_CORE=y). Not
tested at runtime.
v1: https://lore.kernel.org/rust-for-linux/20261114200703.33229-1-mbkchelofficialdev@gmail.com/
rust/kernel/gpu/buddy.rs | 26 +++++++++-----------------
1 file changed, 9 insertions(+), 17 deletions(-)
diff --git a/rust/kernel/gpu/buddy.rs b/rust/kernel/gpu/buddy.rs
index 691bb4062..a85530895 100644
--- a/rust/kernel/gpu/buddy.rs
+++ b/rust/kernel/gpu/buddy.rs
@@ -313,16 +313,11 @@ pub struct GpuBuddyParams {
///
/// # Invariants
///
-/// The inner [`Opaque`] contains an initialized buddy allocator.
+/// The `lock` field contains an initialized buddy allocator wrapped in `Opaque`.
#[pin_data(PinnedDrop)]
struct GpuBuddyInner {
#[pin]
- inner: Opaque<bindings::gpu_buddy>,
-
- // TODO: Replace `Mutex<()>` with `Mutex<Opaque<..>>` once `Mutex::new()`
- // accepts `impl PinInit<T>`.
- #[pin]
- lock: Mutex<()>,
+ lock: Mutex<Opaque<bindings::gpu_buddy>>,
/// Cached creation parameters (do not change after init).
params: GpuBuddyParams,
}
@@ -338,14 +333,13 @@ fn new(params: GpuBuddyParams) -> impl PinInit<Self, Error> {
// `gpu_buddy_*` APIs. `try_pin_init!` only completes if all fields succeed,
// so the invariant holds when construction finishes.
try_pin_init!(Self {
- inner <- Opaque::try_ffi_init(|ptr| {
+ lock <- new_mutex!(Opaque::try_ffi_init(|ptr| {
// SAFETY: `ptr` points to valid uninitialized memory from the pin-init
// infrastructure. `gpu_buddy_init` will initialize the structure.
to_result(unsafe {
bindings::gpu_buddy_init(ptr, size, chunk_size.as_usize() as u64)
})
- }),
- lock <- new_mutex!(()),
+ })),
params,
})
}
@@ -353,7 +347,6 @@ fn new(params: GpuBuddyParams) -> impl PinInit<Self, Error> {
/// Lock the mutex and return a guard for accessing the allocator.
fn lock(&self) -> GpuBuddyGuard<'_> {
GpuBuddyGuard {
- inner: self,
_guard: self.lock.lock(),
}
}
@@ -364,7 +357,7 @@ impl PinnedDrop for GpuBuddyInner {
fn drop(self: Pin<&mut Self>) {
let guard = self.lock();
- // SAFETY: Per the type invariant, `inner` contains an initialized
+ // SAFETY: Per the type invariant, `lock` contains an initialized
// allocator. `guard` provides exclusive access.
unsafe { bindings::gpu_buddy_fini(guard.as_raw()) };
}
@@ -381,14 +374,13 @@ unsafe impl Sync for GpuBuddyInner {}
///
/// The `_guard` holds the lock for the duration of this guard's lifetime.
struct GpuBuddyGuard<'a> {
- inner: &'a GpuBuddyInner,
- _guard: MutexGuard<'a, ()>,
+ _guard: MutexGuard<'a, Opaque<bindings::gpu_buddy>>,
}
impl GpuBuddyGuard<'_> {
/// Get a raw pointer to the underlying C `gpu_buddy` structure.
fn as_raw(&self) -> *mut bindings::gpu_buddy {
- self.inner.inner.get()
+ self._guard.get()
}
}
@@ -428,7 +420,7 @@ pub fn size(&self) -> u64 {
pub fn avail(&self) -> u64 {
let guard = self.0.lock();
- // SAFETY: Per the type invariant, `inner` contains an initialized allocator.
+ // SAFETY: Per the type invariant, `lock` contains an initialized allocator.
// `guard` provides exclusive access.
unsafe { (*guard.as_raw()).avail }
}
@@ -463,7 +455,7 @@ pub fn alloc_blocks(
// Lock while allocating to serialize with concurrent frees.
let guard = buddy.lock();
- // SAFETY: Per the type invariant, `inner` contains an initialized
+ // SAFETY: Per the type invariant, `lock` contains an initialized
// allocator. `guard` provides exclusive access.
to_result(unsafe {
bindings::gpu_buddy_alloc_blocks(
--
2.56.1
reply other threads:[~2026-11-19 11:19 UTC|newest]
Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20261119111850.33270-1-mbkchelofficialdev@gmail.com \
--to=mbkchelofficialdev@gmail.com \
--cc=a.hindborg@kernel.org \
--cc=acourbot@nvidia.com \
--cc=aliceryhl@google.com \
--cc=arunpravin.paneerselvam@amd.com \
--cc=bjorn3_gh@protonmail.com \
--cc=boqun@kernel.org \
--cc=dakr@kernel.org \
--cc=daniel.almeida@collabora.com \
--cc=dri-devel@lists.freedesktop.org \
--cc=gary@garyguo.net \
--cc=jfern@kernel.org \
--cc=lossin@kernel.org \
--cc=matthew.auld@intel.com \
--cc=ojeda@kernel.org \
--cc=rust-for-linux@vger.kernel.org \
--cc=tamird@kernel.org \
--cc=tmgross@umich.edu \
--cc=work@onurozkan.dev \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox