linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Matthew Maurer <mmaurer@google.com>
To: "Miguel Ojeda" <ojeda@kernel.org>,
	"Alex Gaynor" <alex.gaynor@gmail.com>,
	"Boqun Feng" <boqun.feng@gmail.com>,
	"Gary Guo" <gary@garyguo.net>,
	"Björn Roy Baron" <bjorn3_gh@protonmail.com>,
	"Andreas Hindborg" <a.hindborg@kernel.org>,
	"Alice Ryhl" <aliceryhl@google.com>,
	"Trevor Gross" <tmgross@umich.edu>,
	"Danilo Krummrich" <dakr@kernel.org>,
	"Greg Kroah-Hartman" <gregkh@linuxfoundation.org>,
	"Rafael J. Wysocki" <rafael@kernel.org>,
	"Sami Tolvanen" <samitolvanen@google.com>,
	"Timur Tabi" <ttabi@nvidia.com>,
	"Benno Lossin" <lossin@kernel.org>
Cc: linux-kernel@vger.kernel.org, rust-for-linux@vger.kernel.org,
	 Matthew Maurer <mmaurer@google.com>
Subject: [PATCH v9 3/5] rust: debugfs: Support `PinInit` backing for `File`s.
Date: Wed, 09 Jul 2025 19:09:30 +0000	[thread overview]
Message-ID: <20250709-debugfs-rust-v9-3-92b9eab5a951@google.com> (raw)
In-Reply-To: <20250709-debugfs-rust-v9-0-92b9eab5a951@google.com>

This allows `File`s to own their data, allowing DebugFS files to be
managed in sync with the data that backs them. Because DebugFS files are
intended to actually own data and provide access, `File`s still maintain
the same lifecycle for provided data when `CONFIG_DEBUG_FS` is disabled.

Signed-off-by: Matthew Maurer <mmaurer@google.com>
---
 rust/kernel/debugfs.rs | 149 ++++++++++++++++++++++++++++++++++++++-----------
 1 file changed, 117 insertions(+), 32 deletions(-)

diff --git a/rust/kernel/debugfs.rs b/rust/kernel/debugfs.rs
index e5b6497d1deb67671d22ffd90cd5baa855bb9257..a1a84dd309216f455ae8fe3d3c0fd00f957f82a9 100644
--- a/rust/kernel/debugfs.rs
+++ b/rust/kernel/debugfs.rs
@@ -5,12 +5,13 @@
 //!
 //! C header: [`include/linux/debugfs.h`](srctree/include/linux/debugfs.h)
 
-#[cfg(CONFIG_DEBUG_FS)]
-use crate::prelude::GFP_KERNEL;
+use crate::prelude::*;
 use crate::str::CStr;
 #[cfg(CONFIG_DEBUG_FS)]
 use crate::sync::Arc;
 use core::fmt::Display;
+use core::marker::PhantomPinned;
+use core::ops::Deref;
 
 #[cfg(CONFIG_DEBUG_FS)]
 mod display_file;
@@ -63,40 +64,78 @@ fn create(_name: &CStr, _parent: Option<&Dir>) -> Self {
     }
 
     #[cfg(CONFIG_DEBUG_FS)]
-    fn create_file<T: Display + Sized + Sync>(&self, name: &CStr, data: &'static T) -> File {
-        let Some(parent) = &self.0 else {
-            return File {
+    /// Creates a DebugFS file which will own the data produced by the initializer provided in
+    /// `data`.
+    ///
+    /// # Safety
+    ///
+    /// The provided vtable must be appropriate for implementing a seq_file if provided
+    /// with a private data pointer which provides shared access to a `T`.
+    unsafe fn create_file<'a, T: Sync, E, TI: PinInit<T, E>>(
+        &self,
+        name: &'a CStr,
+        data: TI,
+        vtable: &'static bindings::file_operations,
+    ) -> impl PinInit<File<T>, E> + use<'_, 'a, T, E, TI> {
+        try_pin_init! {
+            File {
                 _entry: Entry::empty(),
+                data <- data,
+                _pin: PhantomPinned,
+            } ? E
+        }
+        .pin_chain(|file| {
+            let Some(parent) = &self.0 else {
+                return Ok(());
             };
-        };
-        // SAFETY:
-        // * `name` is a NUL-terminated C string, living across the call, by `CStr` invariant.
-        // * `parent` is a live `dentry` since we have a reference to it.
-        // * `vtable` is all stock `seq_file` implementations except for `open`.
-        //   `open`'s only requirement beyond what is provided to all open functions is that the
-        //   inode's data pointer must point to a `T` that will outlive it, which we know because
-        //   we have a static reference.
-        let ptr = unsafe {
-            bindings::debugfs_create_file_full(
-                name.as_char_ptr(),
-                0o444,
-                parent.as_ptr(),
-                data as *const _ as *mut _,
-                core::ptr::null(),
-                &<T as display_file::DisplayFile>::VTABLE,
-            )
-        };
 
-        // SAFETY: `debugfs_create_file_full` either returns an error code or a legal
-        // dentry pointer, so `Entry::new` is safe to call here.
-        let entry = unsafe { Entry::new(ptr, Some(parent.clone())) };
+            // SAFETY:
+            // * `name` is a NUL-terminated C string, living across the call, by `CStr` invariant.
+            // * `parent` is a live `dentry` since we have a reference to it.
+            // * Since the file owns the `T` and it is pinned, we can safely assume the pointer
+            //   lives and is valid as long as we are.
+            // * Since the `Entry` will live in the `File`, it will be dropped before the pointer
+            //   is invalidated. Dropping the `Entry` will remove the DebugFS file and avoid
+            //   further access.
+            let ptr = unsafe {
+                bindings::debugfs_create_file_full(
+                    name.as_char_ptr(),
+                    0o444,
+                    parent.as_ptr(),
+                    &file.data as *const _ as *mut c_void,
+                    core::ptr::null(),
+                    vtable,
+                )
+            };
+
+            // SAFETY: `debugfs_create_file_full` either returns an error code or a legal
+            // dentry pointer, so `Entry::new` is safe to call here.
+            *file.entry_mut() = unsafe { Entry::new(ptr, Some(parent.clone())) };
 
-        File { _entry: entry }
+            Ok(())
+        })
     }
 
     #[cfg(not(CONFIG_DEBUG_FS))]
-    fn create_file<T: Display + Sized + Sync>(&self, _name: &CStr, _data: &'static T) -> File {
-        File {}
+    /// Creates a DebugFS file which will own the data produced by the initializer provided in
+    /// `data`.
+    ///
+    /// # Safety
+    ///
+    /// As DebugFS is disabled, this is actually entirely safe. It is marked unsafe for code
+    /// compatibility with the DebugFS-enabled variant.
+    unsafe fn create_file<'a, T: Sync, E, TI: PinInit<T, E>>(
+        &self,
+        _name: &'a CStr,
+        data: TI,
+        _vtable: (),
+    ) -> impl PinInit<File<T>, E> + use<'_, 'a, T, E, TI> {
+        try_pin_init! {
+            File {
+                data <- data,
+                _pin: PhantomPinned,
+            } ? E
+        }
     }
 
     /// Create a DebugFS subdirectory.
@@ -127,8 +166,32 @@ pub fn subdir(&self, name: &CStr) -> Self {
     /// dir.display_file(c_str!("foo"), &200);
     /// // "my_debugfs_dir/foo" now contains the number 200.
     /// ```
-    pub fn display_file<T: Display + Sized + Sync>(&self, name: &CStr, data: &'static T) -> File {
-        self.create_file(name, data)
+    ///
+    /// ```
+    /// # use kernel::c_str;
+    /// # use kernel::debugfs::Dir;
+    /// # use kernel::prelude::*;
+    /// let val = KBox::new(300, GFP_KERNEL)?;
+    /// let dir = Dir::new(c_str!("my_debugfs_dir"));
+    /// dir.display_file(c_str!("foo"), val);
+    /// // "my_debugfs_dir/foo" now contains the number 300.
+    /// # Ok::<(), Error>(())
+    /// ```
+    pub fn display_file<'b, T: Display + Send + Sync, E, TI: PinInit<T, E>>(
+        &self,
+        name: &'b CStr,
+        data: TI,
+    ) -> impl PinInit<File<T>, E> + use<'_, 'b, T, E, TI> {
+        #[cfg(CONFIG_DEBUG_FS)]
+        let vtable = &<T as display_file::DisplayFile>::VTABLE;
+        #[cfg(not(CONFIG_DEBUG_FS))]
+        let vtable = ();
+
+        // SAFETY: `vtable` is all stock `seq_file` implementations except for `open`.
+        // `open`'s only requirement beyond what is provided to all open functions is that the
+        // inode's data pointer must point to a `T` that will outlive it, which is provided by
+        // `create_file`'s safety requirements.
+        unsafe { self.create_file(name, data, vtable) }
     }
 
     /// Create a new directory in DebugFS at the root.
@@ -146,7 +209,29 @@ pub fn new(name: &CStr) -> Self {
 }
 
 /// Handle to a DebugFS file.
-pub struct File {
+#[pin_data]
+pub struct File<T> {
+    // This order is load-bearing for drops - `_entry` must be dropped before `data`.
     #[cfg(CONFIG_DEBUG_FS)]
     _entry: Entry,
+    #[pin]
+    data: T,
+    // Even if `T` is `Unpin`, we still can't allow it to be moved.
+    #[pin]
+    _pin: PhantomPinned,
+}
+
+#[cfg(CONFIG_DEBUG_FS)]
+impl<T> File<T> {
+    fn entry_mut(self: Pin<&mut Self>) -> &mut Entry {
+        // SAFETY: _entry is not structurally pinned
+        unsafe { &mut Pin::into_inner_unchecked(self)._entry }
+    }
+}
+
+impl<T> Deref for File<T> {
+    type Target = T;
+    fn deref(&self) -> &T {
+        &self.data
+    }
 }

-- 
2.50.0.727.gbf7dc18ff4-goog


  parent reply	other threads:[~2025-07-09 19:09 UTC|newest]

Thread overview: 23+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-07-09 19:09 [PATCH v9 0/5] rust: DebugFS Bindings Matthew Maurer
2025-07-09 19:09 ` [PATCH v9 1/5] rust: debugfs: Bind DebugFS directory creation Matthew Maurer
2025-07-09 19:09 ` [PATCH v9 2/5] rust: debugfs: Bind file creation for long-lived Display Matthew Maurer
2025-07-09 19:09 ` Matthew Maurer [this message]
2025-08-19  5:51   ` [PATCH v9 3/5] rust: debugfs: Support `PinInit` backing for `File`s Dirk Behme
2025-08-19 14:33     ` Matthew Maurer
2025-08-19 14:47       ` Miguel Ojeda
2025-08-19 23:22         ` Matthew Maurer
2025-07-09 19:09 ` [PATCH v9 4/5] rust: debugfs: Support format hooks Matthew Maurer
2025-07-09 19:09 ` [PATCH v9 5/5] rust: samples: Add debugfs sample Matthew Maurer
2025-07-09 21:56   ` Danilo Krummrich
2025-07-09 23:35     ` Matthew Maurer
2025-07-09 21:47 ` [PATCH v9 0/5] rust: DebugFS Bindings Danilo Krummrich
2025-07-09 21:53   ` Matthew Maurer
2025-07-09 21:59     ` Danilo Krummrich
2025-07-09 22:04       ` Matthew Maurer
2025-07-09 22:12         ` Danilo Krummrich
2025-07-09 22:21           ` Matthew Maurer
2025-07-09 22:33             ` Danilo Krummrich
2025-07-10  5:27           ` Greg Kroah-Hartman
2025-07-10  9:36             ` Danilo Krummrich
2025-07-10 11:09               ` Benno Lossin
2025-07-10 11:11                 ` Benno Lossin

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=20250709-debugfs-rust-v9-3-92b9eab5a951@google.com \
    --to=mmaurer@google.com \
    --cc=a.hindborg@kernel.org \
    --cc=alex.gaynor@gmail.com \
    --cc=aliceryhl@google.com \
    --cc=bjorn3_gh@protonmail.com \
    --cc=boqun.feng@gmail.com \
    --cc=dakr@kernel.org \
    --cc=gary@garyguo.net \
    --cc=gregkh@linuxfoundation.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=lossin@kernel.org \
    --cc=ojeda@kernel.org \
    --cc=rafael@kernel.org \
    --cc=rust-for-linux@vger.kernel.org \
    --cc=samitolvanen@google.com \
    --cc=tmgross@umich.edu \
    --cc=ttabi@nvidia.com \
    /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;
as well as URLs for NNTP newsgroup(s).