ANDROID: rust_binder: move Rust Binder in preparation for GKI module

This moves the code from the common-modules/binder branch verbatim. The
code is not yet added to the build system as it requires some adjusments
before it can built as a GKI module.

Bug: 388786466
Change-Id: I6b49b1b6ff0e35fbae2b9efc13d6bbde984b5196
Signed-off-by: Alice Ryhl <aliceryhl@google.com>

drivers/android/binder/allocation.rs

@@ -0,0 +1,611 @@
+// SPDX-License-Identifier: GPL-2.0
+
+// Copyright (C) 2024 Google LLC.
+
+use core::mem::{size_of, size_of_val, MaybeUninit};
+use core::ops::Range;
+
+use kernel::{
+    bindings,
+    fs::file::{File, FileDescriptorReservation},
+    prelude::*,
+    sync::Arc,
+    types::{ARef, AsBytes, FromBytes},
+    uaccess::UserSliceReader,
+    uapi,
+};
+
+use crate::{
+    deferred_close::DeferredFdCloser,
+    defs::*,
+    node::{Node, NodeRef},
+    process::Process,
+    DArc,
+};
+
+#[derive(Default)]
+pub(crate) struct AllocationInfo {
+    /// Range within the allocation where we can find the offsets to the object descriptors.
+    pub(crate) offsets: Option<Range<usize>>,
+    /// The target node of the transaction this allocation is associated to.
+    /// Not set for replies.
+    pub(crate) target_node: Option<NodeRef>,
+    /// When this allocation is dropped, call `pending_oneway_finished` on the node.
+    ///
+    /// This is used to serialize oneway transaction on the same node. Binder guarantees that
+    /// oneway transactions to the same node are delivered sequentially in the order they are sent.
+    pub(crate) oneway_node: Option<DArc<Node>>,
+    /// Zero the data in the buffer on free.
+    pub(crate) clear_on_free: bool,
+    /// List of files embedded in this transaction.
+    file_list: FileList,
+}
+
+/// Represents an allocation that the kernel is currently using.
+///
+/// When allocations are idle, the range allocator holds the data related to them.
+///
+/// # Invariants
+///
+/// This allocation corresponds to an allocation in the range allocator, so the relevant pages are
+/// marked in use in the page range.
+pub(crate) struct Allocation {
+    pub(crate) offset: usize,
+    size: usize,
+    pub(crate) ptr: usize,
+    pub(crate) process: Arc<Process>,
+    allocation_info: Option<AllocationInfo>,
+    free_on_drop: bool,
+    pub(crate) oneway_spam_detected: bool,
+    #[allow(dead_code)]
+    pub(crate) debug_id: usize,
+}
+
+impl Allocation {
+    pub(crate) fn new(
+        process: Arc<Process>,
+        debug_id: usize,
+        offset: usize,
+        size: usize,
+        ptr: usize,
+        oneway_spam_detected: bool,
+    ) -> Self {
+        Self {
+            process,
+            offset,
+            size,
+            ptr,
+            debug_id,
+            oneway_spam_detected,
+            allocation_info: None,
+            free_on_drop: true,
+        }
+    }
+
+    fn size_check(&self, offset: usize, size: usize) -> Result {
+        let overflow_fail = offset.checked_add(size).is_none();
+        let cmp_size_fail = offset.wrapping_add(size) > self.size;
+        if overflow_fail || cmp_size_fail {
+            return Err(EFAULT);
+        }
+        Ok(())
+    }
+
+    pub(crate) fn copy_into(
+        &self,
+        reader: &mut UserSliceReader,
+        offset: usize,
+        size: usize,
+    ) -> Result {
+        self.size_check(offset, size)?;
+
+        // SAFETY: While this object exists, the range allocator will keep the range allocated, and
+        // in turn, the pages will be marked as in use.
+        unsafe {
+            self.process
+                .pages
+                .copy_from_user_slice(reader, self.offset + offset, size)
+        }
+    }
+
+    pub(crate) fn read<T: FromBytes>(&self, offset: usize) -> Result<T> {
+        self.size_check(offset, size_of::<T>())?;
+
+        // SAFETY: While this object exists, the range allocator will keep the range allocated, and
+        // in turn, the pages will be marked as in use.
+        unsafe { self.process.pages.read(self.offset + offset) }
+    }
+
+    pub(crate) fn write<T: ?Sized>(&self, offset: usize, obj: &T) -> Result {
+        self.size_check(offset, size_of_val::<T>(obj))?;
+
+        // SAFETY: While this object exists, the range allocator will keep the range allocated, and
+        // in turn, the pages will be marked as in use.
+        unsafe { self.process.pages.write(self.offset + offset, obj) }
+    }
+
+    pub(crate) fn fill_zero(&self) -> Result {
+        // SAFETY: While this object exists, the range allocator will keep the range allocated, and
+        // in turn, the pages will be marked as in use.
+        unsafe { self.process.pages.fill_zero(self.offset, self.size) }
+    }
+
+    pub(crate) fn keep_alive(mut self) {
+        self.process
+            .buffer_make_freeable(self.offset, self.allocation_info.take());
+        self.free_on_drop = false;
+    }
+
+    pub(crate) fn set_info(&mut self, info: AllocationInfo) {
+        self.allocation_info = Some(info);
+    }
+
+    pub(crate) fn get_or_init_info(&mut self) -> &mut AllocationInfo {
+        self.allocation_info.get_or_insert_with(Default::default)
+    }
+
+    pub(crate) fn set_info_offsets(&mut self, offsets: Range<usize>) {
+        self.get_or_init_info().offsets = Some(offsets);
+    }
+
+    pub(crate) fn set_info_oneway_node(&mut self, oneway_node: DArc<Node>) {
+        self.get_or_init_info().oneway_node = Some(oneway_node);
+    }
+
+    pub(crate) fn set_info_clear_on_drop(&mut self) {
+        self.get_or_init_info().clear_on_free = true;
+    }
+
+    pub(crate) fn set_info_target_node(&mut self, target_node: NodeRef) {
+        self.get_or_init_info().target_node = Some(target_node);
+    }
+
+    /// Reserve enough space to push at least `num_fds` fds.
+    pub(crate) fn info_add_fd_reserve(&mut self, num_fds: usize) -> Result {
+        self.get_or_init_info()
+            .file_list
+            .files_to_translate
+            .reserve(num_fds, GFP_KERNEL)?;
+
+        Ok(())
+    }
+
+    pub(crate) fn info_add_fd(
+        &mut self,
+        file: ARef<File>,
+        buffer_offset: usize,
+        close_on_free: bool,
+    ) -> Result {
+        self.get_or_init_info().file_list.files_to_translate.push(
+            FileEntry {
+                file,
+                buffer_offset,
+                close_on_free,
+            },
+            GFP_KERNEL,
+        )?;
+
+        Ok(())
+    }
+
+    pub(crate) fn set_info_close_on_free(&mut self, cof: FdsCloseOnFree) {
+        self.get_or_init_info().file_list.close_on_free = cof.0;
+    }
+
+    pub(crate) fn translate_fds(&mut self) -> Result<TranslatedFds> {
+        let file_list = match self.allocation_info.as_mut() {
+            Some(info) => &mut info.file_list,
+            None => return Ok(TranslatedFds::new()),
+        };
+
+        let files = core::mem::take(&mut file_list.files_to_translate);
+
+        let num_close_on_free = files.iter().filter(|entry| entry.close_on_free).count();
+        let mut close_on_free = Vec::with_capacity(num_close_on_free, GFP_KERNEL)?;
+
+        let mut reservations = Vec::with_capacity(files.len(), GFP_KERNEL)?;
+        for file_info in files {
+            let res = FileDescriptorReservation::get_unused_fd_flags(bindings::O_CLOEXEC)?;
+            let fd = res.reserved_fd();
+            self.write::<u32>(file_info.buffer_offset, &fd)?;
+            crate::trace::trace_transaction_fd_recv(self.debug_id, fd, file_info.buffer_offset);
+
+            reservations.push(
+                Reservation {
+                    res,
+                    file: file_info.file,
+                },
+                GFP_KERNEL,
+            )?;
+            if file_info.close_on_free {
+                close_on_free.push(fd, GFP_KERNEL)?;
+            }
+        }
+
+        Ok(TranslatedFds {
+            reservations,
+            close_on_free: FdsCloseOnFree(close_on_free),
+        })
+    }
+
+    /// Should the looper return to userspace when freeing this allocation?
+    pub(crate) fn looper_need_return_on_free(&self) -> bool {
+        // Closing fds involves pushing task_work for execution when we return to userspace. Hence,
+        // we should return to userspace asap if we are closing fds.
+        match self.allocation_info {
+            Some(ref info) => !info.file_list.close_on_free.is_empty(),
+            None => false,
+        }
+    }
+}
+
+impl Drop for Allocation {
+    fn drop(&mut self) {
+        if !self.free_on_drop {
+            return;
+        }
+
+        if let Some(mut info) = self.allocation_info.take() {
+            if let Some(oneway_node) = info.oneway_node.as_ref() {
+                oneway_node.pending_oneway_finished();
+            }
+
+            info.target_node = None;
+
+            if let Some(offsets) = info.offsets.clone() {
+                let view = AllocationView::new(self, offsets.start);
+                for i in offsets.step_by(size_of::<usize>()) {
+                    if view.cleanup_object(i).is_err() {
+                        pr_warn!("Error cleaning up object at offset {}\n", i)
+                    }
+                }
+            }
+
+            for &fd in &info.file_list.close_on_free {
+                let closer = match DeferredFdCloser::new(GFP_KERNEL) {
+                    Ok(closer) => closer,
+                    Err(kernel::alloc::AllocError) => {
+                        // Ignore allocation failures.
+                        break;
+                    }
+                };
+
+                // Here, we ignore errors. The operation can fail if the fd is not valid, or if the
+                // method is called from a kthread. However, this is always called from a syscall,
+                // so the latter case cannot happen, and we don't care about the first case.
+                let _ = closer.close_fd(fd);
+            }
+
+            if info.clear_on_free {
+                if let Err(e) = self.fill_zero() {
+                    pr_warn!("Failed to clear data on free: {:?}", e);
+                }
+            }
+        }
+
+        self.process.buffer_raw_free(self.ptr);
+    }
+}
+
+/// A wrapper around `Allocation` that is being created.
+///
+/// If the allocation is destroyed while wrapped in this wrapper, then the allocation will be
+/// considered to be part of a failed transaction. Successful transactions avoid that by calling
+/// `success`, which skips the destructor.
+#[repr(transparent)]
+pub(crate) struct NewAllocation(pub(crate) Allocation);
+
+impl NewAllocation {
+    pub(crate) fn success(self) -> Allocation {
+        // This skips the destructor.
+        //
+        // SAFETY: This type is `#[repr(transparent)]`, so the layout matches.
+        unsafe { core::mem::transmute(self) }
+    }
+}
+
+impl core::ops::Deref for NewAllocation {
+    type Target = Allocation;
+    fn deref(&self) -> &Allocation {
+        &self.0
+    }
+}
+
+impl core::ops::DerefMut for NewAllocation {
+    fn deref_mut(&mut self) -> &mut Allocation {
+        &mut self.0
+    }
+}
+
+impl Drop for NewAllocation {
+    fn drop(&mut self) {
+        crate::trace::trace_transaction_failed_buffer_release(self.debug_id);
+    }
+}
+
+/// A view into the beginning of an allocation.
+///
+/// All attempts to read or write outside of the view will fail. To intentionally access outside of
+/// this view, use the `alloc` field of this struct directly.
+pub(crate) struct AllocationView<'a> {
+    pub(crate) alloc: &'a mut Allocation,
+    limit: usize,
+}
+
+impl<'a> AllocationView<'a> {
+    pub(crate) fn new(alloc: &'a mut Allocation, limit: usize) -> Self {
+        AllocationView { alloc, limit }
+    }
+
+    pub(crate) fn read<T: FromBytes>(&self, offset: usize) -> Result<T> {
+        if offset.checked_add(size_of::<T>()).ok_or(EINVAL)? > self.limit {
+            return Err(EINVAL);
+        }
+        self.alloc.read(offset)
+    }
+
+    pub(crate) fn write<T: AsBytes>(&self, offset: usize, obj: &T) -> Result {
+        if offset.checked_add(size_of::<T>()).ok_or(EINVAL)? > self.limit {
+            return Err(EINVAL);
+        }
+        self.alloc.write(offset, obj)
+    }
+
+    pub(crate) fn copy_into(
+        &self,
+        reader: &mut UserSliceReader,
+        offset: usize,
+        size: usize,
+    ) -> Result {
+        if offset.checked_add(size).ok_or(EINVAL)? > self.limit {
+            return Err(EINVAL);
+        }
+        self.alloc.copy_into(reader, offset, size)
+    }
+
+    pub(crate) fn transfer_binder_object(
+        &self,
+        offset: usize,
+        obj: &uapi::flat_binder_object,
+        strong: bool,
+        node_ref: NodeRef,
+    ) -> Result {
+        let mut newobj = FlatBinderObject::default();
+        let node = node_ref.node.clone();
+        if Arc::ptr_eq(&node_ref.node.owner, &self.alloc.process) {
+            // The receiving process is the owner of the node, so send it a binder object (instead
+            // of a handle).
+            let (ptr, cookie) = node.get_id();
+            newobj.hdr.type_ = if strong {
+                BINDER_TYPE_BINDER
+            } else {
+                BINDER_TYPE_WEAK_BINDER
+            };
+            newobj.flags = obj.flags;
+            newobj.__bindgen_anon_1.binder = ptr as _;
+            newobj.cookie = cookie as _;
+            self.write(offset, &newobj)?;
+            // Increment the user ref count on the node. It will be decremented as part of the
+            // destruction of the buffer, when we see a binder or weak-binder object.
+            node.update_refcount(true, 1, strong);
+        } else {
+            // The receiving process is different from the owner, so we need to insert a handle to
+            // the binder object.
+            let handle = self
+                .alloc
+                .process
+                .as_arc_borrow()
+                .insert_or_update_handle(node_ref, false)?;
+            newobj.hdr.type_ = if strong {
+                BINDER_TYPE_HANDLE
+            } else {
+                BINDER_TYPE_WEAK_HANDLE
+            };
+            newobj.flags = obj.flags;
+            newobj.__bindgen_anon_1.handle = handle;
+            if self.write(offset, &newobj).is_err() {
+                // Decrement ref count on the handle we just created.
+                let _ = self
+                    .alloc
+                    .process
+                    .as_arc_borrow()
+                    .update_ref(handle, false, strong);
+                return Err(EINVAL);
+            }
+        }
+
+        crate::trace::trace_transaction_node_send(self.alloc.debug_id, &node, obj, &newobj);
+
+        Ok(())
+    }
+
+    fn cleanup_object(&self, index_offset: usize) -> Result {
+        let offset = self.alloc.read(index_offset)?;
+        let header = self.read::<BinderObjectHeader>(offset)?;
+        match header.type_ {
+            BINDER_TYPE_WEAK_BINDER | BINDER_TYPE_BINDER => {
+                let obj = self.read::<FlatBinderObject>(offset)?;
+                let strong = header.type_ == BINDER_TYPE_BINDER;
+                // SAFETY: The type is `BINDER_TYPE_{WEAK_}BINDER`, so the `binder` field is
+                // populated.
+                let ptr = unsafe { obj.__bindgen_anon_1.binder };
+                let cookie = obj.cookie;
+                self.alloc.process.update_node(ptr, cookie, strong);
+                Ok(())
+            }
+            BINDER_TYPE_WEAK_HANDLE | BINDER_TYPE_HANDLE => {
+                let obj = self.read::<FlatBinderObject>(offset)?;
+                let strong = header.type_ == BINDER_TYPE_HANDLE;
+                // SAFETY: The type is `BINDER_TYPE_{WEAK_}HANDLE`, so the `handle` field is
+                // populated.
+                let handle = unsafe { obj.__bindgen_anon_1.handle };
+                self.alloc
+                    .process
+                    .as_arc_borrow()
+                    .update_ref(handle, false, strong)
+            }
+            _ => Ok(()),
+        }
+    }
+}
+
+/// A binder object as it is serialized.
+///
+/// # Invariants
+///
+/// All bytes must be initialized, and the value of `self.hdr.type_` must be one of the allowed
+/// types.
+#[repr(C)]
+pub(crate) union BinderObject {
+    hdr: uapi::binder_object_header,
+    fbo: uapi::flat_binder_object,
+    fdo: uapi::binder_fd_object,
+    bbo: uapi::binder_buffer_object,
+    fdao: uapi::binder_fd_array_object,
+}
+
+/// A view into a `BinderObject` that can be used in a match statement.
+pub(crate) enum BinderObjectRef<'a> {
+    Binder(&'a mut uapi::flat_binder_object),
+    Handle(&'a mut uapi::flat_binder_object),
+    Fd(&'a mut uapi::binder_fd_object),
+    Ptr(&'a mut uapi::binder_buffer_object),
+    Fda(&'a mut uapi::binder_fd_array_object),
+}
+
+impl BinderObject {
+    pub(crate) fn read_from(reader: &mut UserSliceReader) -> Result<BinderObject> {
+        let object = Self::read_from_inner(|slice| {
+            let read_len = usize::min(slice.len(), reader.len());
+            reader.clone_reader().read_slice(&mut slice[..read_len])?;
+            Ok(())
+        })?;
+
+        // If we used a object type smaller than the largest object size, then we've read more
+        // bytes than we needed to. However, we used `.clone_reader()` to avoid advancing the
+        // original reader. Now, we call `skip` so that the caller's reader is advanced by the
+        // right amount.
+        //
+        // The `skip` call fails if the reader doesn't have `size` bytes available. This could
+        // happen if the type header corresponds to an object type that is larger than the rest of
+        // the reader.
+        //
+        // Any extra bytes beyond the size of the object are inaccessible after this call, so
+        // reading them again from the `reader` later does not result in TOCTOU bugs.
+        reader.skip(object.size())?;
+
+        Ok(object)
+    }
+
+    /// Use the provided reader closure to construct a `BinderObject`.
+    ///
+    /// The closure should write the bytes for the object into the provided slice.
+    pub(crate) fn read_from_inner<R>(reader: R) -> Result<BinderObject>
+    where
+        R: FnOnce(&mut [u8; size_of::<BinderObject>()]) -> Result<()>,
+    {
+        let mut obj = MaybeUninit::<BinderObject>::zeroed();
+
+        // SAFETY: The lengths of `BinderObject` and `[u8; size_of::<BinderObject>()]` are equal,
+        // and the byte array has an alignment requirement of one, so the pointer cast is okay.
+        // Additionally, `obj` was initialized to zeros, so the byte array will not be
+        // uninitialized.
+        (reader)(unsafe { &mut *obj.as_mut_ptr().cast() })?;
+
+        // SAFETY: The entire object is initialized, so accessing this field is safe.
+        let type_ = unsafe { obj.assume_init_ref().hdr.type_ };
+        if Self::type_to_size(type_).is_none() {
+            // The value of `obj.hdr_type_` was invalid.
+            return Err(EINVAL);
+        }
+
+        // SAFETY: All bytes are initialized (since we zeroed them at the start) and we checked
+        // that `self.hdr.type_` is one of the allowed types, so the type invariants are satisfied.
+        unsafe { Ok(obj.assume_init()) }
+    }
+
+    pub(crate) fn as_ref(&mut self) -> BinderObjectRef<'_> {
+        use BinderObjectRef::*;
+        // SAFETY: The constructor ensures that all bytes of `self` are initialized, and all
+        // variants of this union accept all initialized bit patterns.
+        unsafe {
+            match self.hdr.type_ {
+                BINDER_TYPE_WEAK_BINDER | BINDER_TYPE_BINDER => Binder(&mut self.fbo),
+                BINDER_TYPE_WEAK_HANDLE | BINDER_TYPE_HANDLE => Handle(&mut self.fbo),
+                BINDER_TYPE_FD => Fd(&mut self.fdo),
+                BINDER_TYPE_PTR => Ptr(&mut self.bbo),
+                BINDER_TYPE_FDA => Fda(&mut self.fdao),
+                // SAFETY: By the type invariant, the value of `self.hdr.type_` cannot have any
+                // other value than the ones checked above.
+                _ => core::hint::unreachable_unchecked(),
+            }
+        }
+    }
+
+    pub(crate) fn size(&self) -> usize {
+        // SAFETY: The entire object is initialized, so accessing this field is safe.
+        let type_ = unsafe { self.hdr.type_ };
+
+        // SAFETY: The type invariants guarantee that the type field is correct.
+        unsafe { Self::type_to_size(type_).unwrap_unchecked() }
+    }
+
+    fn type_to_size(type_: u32) -> Option<usize> {
+        match type_ {
+            BINDER_TYPE_WEAK_BINDER => Some(size_of::<uapi::flat_binder_object>()),
+            BINDER_TYPE_BINDER => Some(size_of::<uapi::flat_binder_object>()),
+            BINDER_TYPE_WEAK_HANDLE => Some(size_of::<uapi::flat_binder_object>()),
+            BINDER_TYPE_HANDLE => Some(size_of::<uapi::flat_binder_object>()),
+            BINDER_TYPE_FD => Some(size_of::<uapi::binder_fd_object>()),
+            BINDER_TYPE_PTR => Some(size_of::<uapi::binder_buffer_object>()),
+            BINDER_TYPE_FDA => Some(size_of::<uapi::binder_fd_array_object>()),
+            _ => None,
+        }
+    }
+}
+
+#[derive(Default)]
+struct FileList {
+    files_to_translate: Vec<FileEntry>,
+    close_on_free: Vec<u32>,
+}
+
+struct FileEntry {
+    /// The file for which a descriptor will be created in the recipient process.
+    file: ARef<File>,
+    /// The offset in the buffer where the file descriptor is stored.
+    buffer_offset: usize,
+    /// Whether this fd should be closed when the allocation is freed.
+    close_on_free: bool,
+}
+
+pub(crate) struct TranslatedFds {
+    reservations: Vec<Reservation>,
+    /// If commit is called, then these fds should be closed. (If commit is not called, then they
+    /// shouldn't be closed.)
+    close_on_free: FdsCloseOnFree,
+}
+
+struct Reservation {
+    res: FileDescriptorReservation,
+    file: ARef<File>,
+}
+
+impl TranslatedFds {
+    pub(crate) fn new() -> Self {
+        Self {
+            reservations: Vec::new(),
+            close_on_free: FdsCloseOnFree(Vec::new()),
+        }
+    }
+
+    pub(crate) fn commit(self) -> FdsCloseOnFree {
+        for entry in self.reservations {
+            entry.res.fd_install(entry.file);
+        }
+
+        self.close_on_free
+    }
+}
+
+pub(crate) struct FdsCloseOnFree(Vec<u32>);

drivers/android/binder/context.rs

@@ -0,0 +1,180 @@
+// SPDX-License-Identifier: GPL-2.0
+
+// Copyright (C) 2024 Google LLC.
+
+use kernel::{
+    error::Error,
+    list::{List, ListArc, ListLinks},
+    prelude::*,
+    security,
+    str::{CStr, CString},
+    sync::{Arc, Mutex},
+    task::Kuid,
+};
+
+use crate::{error::BinderError, node::NodeRef, process::Process};
+
+kernel::sync::global_lock! {
+    // SAFETY: We call `init` in the module initializer, so it's initialized before first use.
+    pub(crate) unsafe(uninit) static CONTEXTS: Mutex<ContextList> = ContextList {
+        list: List::new(),
+    };
+}
+
+pub(crate) struct ContextList {
+    list: List<Context>,
+}
+
+pub(crate) fn get_all_contexts() -> Result<Vec<Arc<Context>>> {
+    let lock = CONTEXTS.lock();
+
+    let count = lock.list.iter().count();
+
+    let mut ctxs = Vec::with_capacity(count, GFP_KERNEL)?;
+    for ctx in &lock.list {
+        ctxs.push(Arc::from(ctx), GFP_KERNEL)?;
+    }
+    Ok(ctxs)
+}
+
+/// This struct keeps track of the processes using this context, and which process is the context
+/// manager.
+struct Manager {
+    node: Option<NodeRef>,
+    uid: Option<Kuid>,
+    all_procs: List<Process>,
+}
+
+/// There is one context per binder file (/dev/binder, /dev/hwbinder, etc)
+#[pin_data]
+pub(crate) struct Context {
+    #[pin]
+    manager: Mutex<Manager>,
+    pub(crate) name: CString,
+    #[pin]
+    links: ListLinks,
+}
+
+kernel::list::impl_has_list_links! {
+    impl HasListLinks<0> for Context { self.links }
+}
+kernel::list::impl_list_arc_safe! {
+    impl ListArcSafe<0> for Context { untracked; }
+}
+kernel::list::impl_list_item! {
+    impl ListItem<0> for Context {
+        using ListLinks;
+    }
+}
+
+impl Context {
+    pub(crate) fn new(name: &CStr) -> Result<Arc<Self>> {
+        let name = CString::try_from(name)?;
+        let list_ctx = ListArc::pin_init::<Error>(
+            try_pin_init!(Context {
+                name,
+                links <- ListLinks::new(),
+                manager <- kernel::new_mutex!(Manager {
+                    all_procs: List::new(),
+                    node: None,
+                    uid: None,
+                }, "Context::manager"),
+            }),
+            GFP_KERNEL,
+        )?;
+
+        let ctx = list_ctx.clone_arc();
+        CONTEXTS.lock().list.push_back(list_ctx);
+
+        Ok(ctx)
+    }
+
+    /// Called when the file for this context is unlinked.
+    ///
+    /// No-op if called twice.
+    pub(crate) fn deregister(&self) {
+        // SAFETY: We never add the context to any other linked list than this one, so it is either
+        // in this list, or not in any list.
+        unsafe { CONTEXTS.lock().list.remove(self) };
+    }
+
+    pub(crate) fn register_process(self: &Arc<Self>, proc: ListArc<Process>) {
+        if !Arc::ptr_eq(self, &proc.ctx) {
+            pr_err!("Context::register_process called on the wrong context.");
+            return;
+        }
+        self.manager.lock().all_procs.push_back(proc);
+    }
+
+    pub(crate) fn deregister_process(self: &Arc<Self>, proc: &Process) {
+        if !Arc::ptr_eq(self, &proc.ctx) {
+            pr_err!("Context::deregister_process called on the wrong context.");
+            return;
+        }
+        // SAFETY: We just checked that this is the right list.
+        unsafe { self.manager.lock().all_procs.remove(proc) };
+    }
+
+    pub(crate) fn set_manager_node(&self, node_ref: NodeRef) -> Result {
+        let mut manager = self.manager.lock();
+        if manager.node.is_some() {
+            pr_warn!("BINDER_SET_CONTEXT_MGR already set");
+            return Err(EBUSY);
+        }
+        security::binder_set_context_mgr(&node_ref.node.owner.cred)?;
+
+        // If the context manager has been set before, ensure that we use the same euid.
+        let caller_uid = Kuid::current_euid();
+        if let Some(ref uid) = manager.uid {
+            if *uid != caller_uid {
+                return Err(EPERM);
+            }
+        }
+
+        manager.node = Some(node_ref);
+        manager.uid = Some(caller_uid);
+        Ok(())
+    }
+
+    pub(crate) fn unset_manager_node(&self) {
+        let node_ref = self.manager.lock().node.take();
+        drop(node_ref);
+    }
+
+    pub(crate) fn get_manager_node(&self, strong: bool) -> Result<NodeRef, BinderError> {
+        self.manager
+            .lock()
+            .node
+            .as_ref()
+            .ok_or_else(BinderError::new_dead)?
+            .clone(strong)
+            .map_err(BinderError::from)
+    }
+
+    pub(crate) fn for_each_proc<F>(&self, mut func: F)
+    where
+        F: FnMut(&Process),
+    {
+        let lock = self.manager.lock();
+        for proc in &lock.all_procs {
+            func(&proc);
+        }
+    }
+
+    pub(crate) fn get_all_procs(&self) -> Result<Vec<Arc<Process>>> {
+        let lock = self.manager.lock();
+        let count = lock.all_procs.iter().count();
+
+        let mut procs = Vec::with_capacity(count, GFP_KERNEL)?;
+        for proc in &lock.all_procs {
+            procs.push(Arc::from(proc), GFP_KERNEL)?;
+        }
+        Ok(procs)
+    }
+
+    pub(crate) fn get_procs_with_pid(&self, pid: i32) -> Result<Vec<Arc<Process>>> {
+        let mut procs = self.get_all_procs()?;
+        procs.retain(|proc| proc.task.pid() == pid);
+        Ok(procs)
+    }
+}

drivers/android/binder/deferred_close.rs

@@ -0,0 +1,202 @@
+// SPDX-License-Identifier: GPL-2.0
+
+// Copyright (C) 2024 Google LLC.
+
+//! Logic for closing files in a deferred manner.
+//!
+//! This file could make sense to have in `kernel::fs`, but it was rejected for being too
+//! Binder-specific.
+
+use core::mem::MaybeUninit;
+use kernel::{
+    alloc::{AllocError, Flags},
+    bindings,
+    prelude::*,
+};
+
+/// Helper used for closing file descriptors in a way that is safe even if the file is currently
+/// held using `fdget`.
+///
+/// Additional motivation can be found in commit 80cd795630d6 ("binder: fix use-after-free due to
+/// ksys_close() during fdget()") and in the comments on `binder_do_fd_close`.
+pub(crate) struct DeferredFdCloser {
+    inner: Box<DeferredFdCloserInner>,
+}
+
+/// SAFETY: This just holds an allocation with no real content, so there's no safety issue with
+/// moving it across threads.
+unsafe impl Send for DeferredFdCloser {}
+unsafe impl Sync for DeferredFdCloser {}
+
+/// # Invariants
+///
+/// If the `file` pointer is non-null, then it points at a `struct file` and owns a refcount to
+/// that file.
+#[repr(C)]
+struct DeferredFdCloserInner {
+    twork: MaybeUninit<bindings::callback_head>,
+    file: *mut bindings::file,
+}
+
+impl DeferredFdCloser {
+    /// Create a new [`DeferredFdCloser`].
+    pub(crate) fn new(flags: Flags) -> Result<Self, AllocError> {
+        Ok(Self {
+            // INVARIANT: The `file` pointer is null, so the type invariant does not apply.
+            inner: Box::new(
+                DeferredFdCloserInner {
+                    twork: MaybeUninit::uninit(),
+                    file: core::ptr::null_mut(),
+                },
+                flags,
+            )?,
+        })
+    }
+
+    /// Schedule a task work that closes the file descriptor when this task returns to userspace.
+    ///
+    /// Fails if this is called from a context where we cannot run work when returning to
+    /// userspace. (E.g., from a kthread.)
+    pub(crate) fn close_fd(self, fd: u32) -> Result<(), DeferredFdCloseError> {
+        use bindings::task_work_notify_mode_TWA_RESUME as TWA_RESUME;
+
+        // In this method, we schedule the task work before closing the file. This is because
+        // scheduling a task work is fallible, and we need to know whether it will fail before we
+        // attempt to close the file.
+
+        // Task works are not available on kthreads.
+        let current = kernel::current!();
+
+        // Check if this is a kthread.
+        // SAFETY: Reading `flags` from a task is always okay.
+        if unsafe { ((*current.as_ptr()).flags & bindings::PF_KTHREAD) != 0 } {
+            return Err(DeferredFdCloseError::TaskWorkUnavailable);
+        }
+
+        // Transfer ownership of the box's allocation to a raw pointer. This disables the
+        // destructor, so we must manually convert it back to a Box to drop it.
+        //
+        // Until we convert it back to a `Box`, there are no aliasing requirements on this
+        // pointer.
+        let inner = Box::into_raw(self.inner);
+
+        // The `callback_head` field is first in the struct, so this cast correctly gives us a
+        // pointer to the field.
+        let callback_head = inner.cast::<bindings::callback_head>();
+        // SAFETY: This pointer offset operation does not go out-of-bounds.
+        let file_field = unsafe { core::ptr::addr_of_mut!((*inner).file) };
+
+        let current = current.as_ptr();
+
+        // SAFETY: This function currently has exclusive access to the `DeferredFdCloserInner`, so
+        // it is okay for us to perform unsynchronized writes to its `callback_head` field.
+        unsafe { bindings::init_task_work(callback_head, Some(Self::do_close_fd)) };
+
+        // SAFETY: This inserts the `DeferredFdCloserInner` into the task workqueue for the current
+        // task. If this operation is successful, then this transfers exclusive ownership of the
+        // `callback_head` field to the C side until it calls `do_close_fd`, and we don't touch or
+        // invalidate the field during that time.
+        //
+        // When the C side calls `do_close_fd`, the safety requirements of that method are
+        // satisfied because when a task work is executed, the callback is given ownership of the
+        // pointer.
+        //
+        // The file pointer is currently null. If it is changed to be non-null before `do_close_fd`
+        // is called, then that change happens due to the write at the end of this function, and
+        // that write has a safety comment that explains why the refcount can be dropped when
+        // `do_close_fd` runs.
+        let res = unsafe { bindings::task_work_add(current, callback_head, TWA_RESUME) };
+
+        if res != 0 {
+            // SAFETY: Scheduling the task work failed, so we still have ownership of the box, so
+            // we may destroy it.
+            unsafe { drop(Box::from_raw(inner)) };
+
+            return Err(DeferredFdCloseError::TaskWorkUnavailable);
+        }
+
+        // This removes the fd from the fd table in `current`. The file is not fully closed until
+        // `filp_close` is called. We are given ownership of one refcount to the file.
+        //
+        // SAFETY: This is safe no matter what `fd` is. If the `fd` is valid (that is, if the
+        // pointer is non-null), then we call `filp_close` on the returned pointer as required by
+        // `file_close_fd`.
+        let file = unsafe { bindings::file_close_fd(fd) };
+        if file.is_null() {
+            // We don't clean up the task work since that might be expensive if the task work queue
+            // is long. Just let it execute and let it clean up for itself.
+            return Err(DeferredFdCloseError::BadFd);
+        }
+
+        // Acquire a second refcount to the file.
+        //
+        // SAFETY: The `file` pointer points at a file with a non-zero refcount.
+        unsafe { bindings::get_file(file) };
+
+        // This method closes the fd, consuming one of our two refcounts. There could be active
+        // light refcounts created from that fd, so we must ensure that the file has a positive
+        // refcount for the duration of those active light refcounts. We do that by holding on to
+        // the second refcount until the current task returns to userspace.
+        //
+        // SAFETY: The `file` pointer is valid. Passing `current->files` as the file table to close
+        // it in is correct, since we just got the `fd` from `file_close_fd` which also uses
+        // `current->files`.
+        //
+        // Note: fl_owner_t is currently a void pointer.
+        unsafe { bindings::filp_close(file, (*current).files as bindings::fl_owner_t) };
+
+        // We update the file pointer that the task work is supposed to fput. This transfers
+        // ownership of our last refcount.
+        //
+        // INVARIANT: This changes the `file` field of a `DeferredFdCloserInner` from null to
+        // non-null. This doesn't break the type invariant for `DeferredFdCloserInner` because we
+        // still own a refcount to the file, so we can pass ownership of that refcount to the
+        // `DeferredFdCloserInner`.
+        //
+        // When `do_close_fd` runs, it must be safe for it to `fput` the refcount. However, this is
+        // the case because all light refcounts that are associated with the fd we closed
+        // previously must be dropped when `do_close_fd`, since light refcounts must be dropped
+        // before returning to userspace.
+        //
+        // SAFETY: Task works are executed on the current thread right before we return to
+        // userspace, so this write is guaranteed to happen before `do_close_fd` is called, which
+        // means that a race is not possible here.
+        unsafe { *file_field = file };
+
+        Ok(())
+    }
+
+    /// # Safety
+    ///
+    /// The provided pointer must point at the `twork` field of a `DeferredFdCloserInner` stored in
+    /// a `Box`, and the caller must pass exclusive ownership of that `Box`. Furthermore, if the
+    /// file pointer is non-null, then it must be okay to release the refcount by calling `fput`.
+    unsafe extern "C" fn do_close_fd(inner: *mut bindings::callback_head) {
+        // SAFETY: The caller just passed us ownership of this box.
+        let inner = unsafe { Box::from_raw(inner.cast::<DeferredFdCloserInner>()) };
+        if !inner.file.is_null() {
+            // SAFETY: By the type invariants, we own a refcount to this file, and the caller
+            // guarantees that dropping the refcount now is okay.
+            unsafe { bindings::fput(inner.file) };
+        }
+        // The allocation is freed when `inner` goes out of scope.
+    }
+}
+
+/// Represents a failure to close an fd in a deferred manner.
+#[derive(Copy, Clone, Debug, Eq, PartialEq)]
+pub(crate) enum DeferredFdCloseError {
+    /// Closing the fd failed because we were unable to schedule a task work.
+    TaskWorkUnavailable,
+    /// Closing the fd failed because the fd does not exist.
+    BadFd,
+}
+
+impl From<DeferredFdCloseError> for Error {
+    fn from(err: DeferredFdCloseError) -> Error {
+        match err {
+            DeferredFdCloseError::TaskWorkUnavailable => ESRCH,
+            DeferredFdCloseError::BadFd => EBADF,
+        }
+    }
+}

drivers/android/binder/defs.rs

@@ -0,0 +1,182 @@
+// SPDX-License-Identifier: GPL-2.0
+
+// Copyright (C) 2024 Google LLC.
+
+use core::mem::MaybeUninit;
+use core::ops::{Deref, DerefMut};
+use kernel::{
+    types::{AsBytes, FromBytes},
+    uapi::{self, *},
+};
+
+macro_rules! pub_no_prefix {
+    ($prefix:ident, $($newname:ident),+ $(,)?) => {
+        $(pub(crate) const $newname: u32 = kernel::macros::concat_idents!($prefix, $newname);)+
+    };
+}
+
+pub_no_prefix!(
+    binder_driver_return_protocol_,
+    BR_TRANSACTION,
+    BR_TRANSACTION_SEC_CTX,
+    BR_REPLY,
+    BR_DEAD_REPLY,
+    BR_FAILED_REPLY,
+    BR_FROZEN_REPLY,
+    BR_NOOP,
+    BR_SPAWN_LOOPER,
+    BR_TRANSACTION_COMPLETE,
+    BR_TRANSACTION_PENDING_FROZEN,
+    BR_ONEWAY_SPAM_SUSPECT,
+    BR_OK,
+    BR_ERROR,
+    BR_INCREFS,
+    BR_ACQUIRE,
+    BR_RELEASE,
+    BR_DECREFS,
+    BR_DEAD_BINDER,
+    BR_CLEAR_DEATH_NOTIFICATION_DONE,
+);
+
+pub_no_prefix!(
+    binder_driver_command_protocol_,
+    BC_TRANSACTION,
+    BC_TRANSACTION_SG,
+    BC_REPLY,
+    BC_REPLY_SG,
+    BC_FREE_BUFFER,
+    BC_ENTER_LOOPER,
+    BC_EXIT_LOOPER,
+    BC_REGISTER_LOOPER,
+    BC_INCREFS,
+    BC_ACQUIRE,
+    BC_RELEASE,
+    BC_DECREFS,
+    BC_INCREFS_DONE,
+    BC_ACQUIRE_DONE,
+    BC_REQUEST_DEATH_NOTIFICATION,
+    BC_CLEAR_DEATH_NOTIFICATION,
+    BC_DEAD_BINDER_DONE,
+);
+
+pub_no_prefix!(
+    flat_binder_object_shifts_,
+    FLAT_BINDER_FLAG_SCHED_POLICY_SHIFT
+);
+
+pub_no_prefix!(
+    flat_binder_object_flags_,
+    FLAT_BINDER_FLAG_ACCEPTS_FDS,
+    FLAT_BINDER_FLAG_INHERIT_RT,
+    FLAT_BINDER_FLAG_PRIORITY_MASK,
+    FLAT_BINDER_FLAG_SCHED_POLICY_MASK,
+    FLAT_BINDER_FLAG_TXN_SECURITY_CTX
+);
+
+pub_no_prefix!(
+    transaction_flags_,
+    TF_ONE_WAY,
+    TF_ACCEPT_FDS,
+    TF_CLEAR_BUF,
+    TF_UPDATE_TXN
+);
+
+pub(crate) use uapi::{
+    BINDER_TYPE_BINDER, BINDER_TYPE_FD, BINDER_TYPE_FDA, BINDER_TYPE_HANDLE, BINDER_TYPE_PTR,
+    BINDER_TYPE_WEAK_BINDER, BINDER_TYPE_WEAK_HANDLE,
+};
+
+macro_rules! decl_wrapper {
+    ($newname:ident, $wrapped:ty) => {
+        // Define a wrapper around the C type. Use `MaybeUninit` to enforce that the value of
+        // padding bytes must be preserved.
+        #[derive(Copy, Clone)]
+        #[repr(transparent)]
+        pub(crate) struct $newname(MaybeUninit<$wrapped>);
+
+        // SAFETY: This macro is only used with types where this is ok.
+        unsafe impl FromBytes for $newname {}
+        unsafe impl AsBytes for $newname {}
+
+        impl Deref for $newname {
+            type Target = $wrapped;
+            fn deref(&self) -> &Self::Target {
+                // SAFETY: We use `MaybeUninit` only to preserve padding. The value must still
+                // always be valid.
+                unsafe { self.0.assume_init_ref() }
+            }
+        }
+
+        impl DerefMut for $newname {
+            fn deref_mut(&mut self) -> &mut Self::Target {
+                // SAFETY: We use `MaybeUninit` only to preserve padding. The value must still
+                // always be valid.
+                unsafe { self.0.assume_init_mut() }
+            }
+        }
+
+        impl Default for $newname {
+            fn default() -> Self {
+                // Create a new value of this type where all bytes (including padding) are zeroed.
+                Self(MaybeUninit::zeroed())
+            }
+        }
+    };
+}
+
+decl_wrapper!(BinderNodeDebugInfo, uapi::binder_node_debug_info);
+decl_wrapper!(BinderNodeInfoForRef, uapi::binder_node_info_for_ref);
+decl_wrapper!(FlatBinderObject, uapi::flat_binder_object);
+decl_wrapper!(BinderFdObject, uapi::binder_fd_object);
+decl_wrapper!(BinderFdArrayObject, uapi::binder_fd_array_object);
+decl_wrapper!(BinderObjectHeader, uapi::binder_object_header);
+decl_wrapper!(BinderBufferObject, uapi::binder_buffer_object);
+decl_wrapper!(BinderTransactionData, uapi::binder_transaction_data);
+decl_wrapper!(
+    BinderTransactionDataSecctx,
+    uapi::binder_transaction_data_secctx
+);
+decl_wrapper!(BinderTransactionDataSg, uapi::binder_transaction_data_sg);
+decl_wrapper!(BinderWriteRead, uapi::binder_write_read);
+decl_wrapper!(BinderVersion, uapi::binder_version);
+decl_wrapper!(BinderFrozenStatusInfo, uapi::binder_frozen_status_info);
+decl_wrapper!(BinderFreezeInfo, uapi::binder_freeze_info);
+decl_wrapper!(ExtendedError, uapi::binder_extended_error);
+
+impl BinderVersion {
+    pub(crate) fn current() -> Self {
+        Self(MaybeUninit::new(uapi::binder_version {
+            protocol_version: BINDER_CURRENT_PROTOCOL_VERSION as _,
+        }))
+    }
+}
+
+impl BinderTransactionData {
+    pub(crate) fn with_buffers_size(self, buffers_size: u64) -> BinderTransactionDataSg {
+        BinderTransactionDataSg(MaybeUninit::new(uapi::binder_transaction_data_sg {
+            transaction_data: *self,
+            buffers_size,
+        }))
+    }
+}
+
+impl BinderTransactionDataSecctx {
+    /// View the inner data as wrapped in `BinderTransactionData`.
+    pub(crate) fn tr_data(&mut self) -> &mut BinderTransactionData {
+        // SAFETY: Transparent wrapper is safe to transmute.
+        unsafe {
+            &mut *(&mut self.transaction_data as *mut uapi::binder_transaction_data
+                as *mut BinderTransactionData)
+        }
+    }
+}
+
+impl ExtendedError {
+    pub(crate) fn new(id: u32, command: u32, param: i32) -> Self {
+        Self(MaybeUninit::new(uapi::binder_extended_error {
+            id,
+            command,
+            param,
+        }))
+    }
+}

drivers/android/binder/error.rs

@@ -0,0 +1,99 @@
+// SPDX-License-Identifier: GPL-2.0
+
+// Copyright (C) 2024 Google LLC.
+
+use kernel::prelude::*;
+
+use crate::defs::*;
+
+pub(crate) type BinderResult<T = ()> = core::result::Result<T, BinderError>;
+
+/// An error that will be returned to userspace via the `BINDER_WRITE_READ` ioctl rather than via
+/// errno.
+pub(crate) struct BinderError {
+    pub(crate) reply: u32,
+    source: Option<Error>,
+}
+
+impl BinderError {
+    pub(crate) fn new_dead() -> Self {
+        Self {
+            reply: BR_DEAD_REPLY,
+            source: None,
+        }
+    }
+
+    pub(crate) fn new_frozen() -> Self {
+        Self {
+            reply: BR_FROZEN_REPLY,
+            source: None,
+        }
+    }
+
+    pub(crate) fn new_frozen_oneway() -> Self {
+        Self {
+            reply: BR_TRANSACTION_PENDING_FROZEN,
+            source: None,
+        }
+    }
+
+    pub(crate) fn is_dead(&self) -> bool {
+        self.reply == BR_DEAD_REPLY
+    }
+
+    pub(crate) fn as_errno(&self) -> core::ffi::c_int {
+        self.source.unwrap_or(EINVAL).to_errno()
+    }
+
+    pub(crate) fn should_pr_warn(&self) -> bool {
+        self.source.is_some()
+    }
+}
+
+/// Convert an errno into a `BinderError` and store the errno used to construct it. The errno
+/// should be stored as the thread's extended error when given to userspace.
+impl From<Error> for BinderError {
+    fn from(source: Error) -> Self {
+        Self {
+            reply: BR_FAILED_REPLY,
+            source: Some(source),
+        }
+    }
+}
+
+impl From<kernel::fs::file::BadFdError> for BinderError {
+    fn from(source: kernel::fs::file::BadFdError) -> Self {
+        BinderError::from(Error::from(source))
+    }
+}
+
+impl From<kernel::alloc::AllocError> for BinderError {
+    fn from(_: kernel::alloc::AllocError) -> Self {
+        Self {
+            reply: BR_FAILED_REPLY,
+            source: Some(ENOMEM),
+        }
+    }
+}
+
+impl core::fmt::Debug for BinderError {
+    fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result {
+        match self.reply {
+            BR_FAILED_REPLY => match self.source.as_ref() {
+                Some(source) => f
+                    .debug_struct("BR_FAILED_REPLY")
+                    .field("source", source)
+                    .finish(),
+                None => f.pad("BR_FAILED_REPLY"),
+            },
+            BR_DEAD_REPLY => f.pad("BR_DEAD_REPLY"),
+            BR_FROZEN_REPLY => f.pad("BR_FROZEN_REPLY"),
+            BR_TRANSACTION_PENDING_FROZEN => f.pad("BR_TRANSACTION_PENDING_FROZEN"),
+            BR_TRANSACTION_COMPLETE => f.pad("BR_TRANSACTION_COMPLETE"),
+            _ => f
+                .debug_struct("BinderError")
+                .field("reply", &self.reply)
+                .finish(),
+        }
+    }
+}

drivers/android/binder/node/wrapper.rs

@@ -0,0 +1,79 @@
+// SPDX-License-Identifier: GPL-2.0
+
+// Copyright (C) 2024 Google LLC.
+
+use kernel::{list::ListArc, prelude::*, seq_file::SeqFile, seq_print, sync::UniqueArc};
+
+use crate::{node::Node, thread::Thread, BinderReturnWriter, DArc, DLArc, DTRWrap, DeliverToRead};
+
+use core::mem::MaybeUninit;
+
+pub(crate) struct CritIncrWrapper {
+    inner: UniqueArc<MaybeUninit<DTRWrap<NodeWrapper>>>,
+}
+
+impl CritIncrWrapper {
+    pub(crate) fn new() -> Result<Self> {
+        Ok(CritIncrWrapper {
+            inner: UniqueArc::new_uninit(GFP_KERNEL)?,
+        })
+    }
+
+    pub(super) fn init(self, node: DArc<Node>) -> DLArc<dyn DeliverToRead> {
+        match self.inner.pin_init_with(DTRWrap::new(NodeWrapper { node })) {
+            Ok(initialized) => ListArc::from(initialized) as _,
+            Err(err) => match err {},
+        }
+    }
+}
+
+struct NodeWrapper {
+    node: DArc<Node>,
+}
+
+kernel::list::impl_list_arc_safe! {
+    impl ListArcSafe<0> for NodeWrapper {
+        untracked;
+    }
+}
+
+impl DeliverToRead for NodeWrapper {
+    fn do_work(
+        self: DArc<Self>,
+        _thread: &Thread,
+        writer: &mut BinderReturnWriter<'_>,
+    ) -> Result<bool> {
+        let node = &self.node;
+        let mut owner_inner = node.owner.inner.lock();
+        let inner = node.inner.access_mut(&mut owner_inner);
+
+        let ds = &mut inner.delivery_state;
+
+        assert!(ds.has_pushed_wrapper);
+        assert!(ds.has_strong_zero2one);
+        ds.has_pushed_wrapper = false;
+        ds.has_strong_zero2one = false;
+
+        node.do_work_locked(writer, owner_inner)
+    }
+
+    fn cancel(self: DArc<Self>) {}
+    fn on_thread_selected(&self, _thread: &Thread) {}
+
+    fn should_sync_wakeup(&self) -> bool {
+        false
+    }
+
+    #[inline(never)]
+    fn debug_print(&self, m: &SeqFile, prefix: &str, _tprefix: &str) -> Result<()> {
+        seq_print!(
+            m,
+            "{}node work {}: u{:016x} c{:016x}\n",
+            prefix,
+            self.node.debug_id,
+            self.node.ptr,
+            self.node.cookie,
+        );
+        Ok(())
+    }
+}

drivers/android/binder/page_range.rs

@@ -0,0 +1,780 @@
+// SPDX-License-Identifier: GPL-2.0
+
+// Copyright (C) 2024 Google LLC.
+
+//! This module has utilities for managing a page range where unused pages may be reclaimed by a
+//! vma shrinker.
+
+// To avoid deadlocks, locks are taken in the order:
+//
+//  1. mmap lock
+//  2. spinlock
+//  3. lru spinlock
+//
+// The shrinker will use trylock methods because it locks them in a different order.
+
+use core::{
+    alloc::Layout,
+    ffi::{c_ulong, c_void},
+    marker::PhantomPinned,
+    mem::{size_of, size_of_val, MaybeUninit},
+    ptr,
+};
+
+use kernel::{
+    bindings,
+    error::Result,
+    mm::{virt, Mm, MmWithUser},
+    new_mutex, new_spinlock,
+    page::{Page, PAGE_SHIFT, PAGE_SIZE},
+    prelude::*,
+    str::CStr,
+    sync::{Mutex, SpinLock},
+    task::Pid,
+    types::ARef,
+    types::{FromBytes, Opaque},
+    uaccess::UserSliceReader,
+};
+
+/// Represents a shrinker that can be registered with the kernel.
+///
+/// Each shrinker can be used by many `ShrinkablePageRange` objects.
+#[repr(C)]
+pub(crate) struct Shrinker {
+    inner: Opaque<*mut bindings::shrinker>,
+    list_lru: Opaque<bindings::list_lru>,
+}
+
+unsafe impl Send for Shrinker {}
+unsafe impl Sync for Shrinker {}
+
+impl Shrinker {
+    /// Create a new shrinker.
+    ///
+    /// # Safety
+    ///
+    /// Before using this shrinker with a `ShrinkablePageRange`, the `register` method must have
+    /// been called exactly once, and it must not have returned an error.
+    pub(crate) const unsafe fn new() -> Self {
+        Self {
+            inner: Opaque::uninit(),
+            list_lru: Opaque::uninit(),
+        }
+    }
+
+    /// Register this shrinker with the kernel.
+    pub(crate) fn register(&'static self, name: &CStr) -> Result<()> {
+        // SAFETY: These fields are not yet used, so it's okay to zero them.
+        unsafe {
+            self.inner.get().write(ptr::null_mut());
+            self.list_lru.get().write_bytes(0, 1);
+        }
+
+        // SAFETY: The field is not yet used, so we can initialize it.
+        let ret = unsafe {
+            bindings::__list_lru_init(self.list_lru.get(), false, ptr::null_mut(), ptr::null_mut())
+        };
+        if ret != 0 {
+            return Err(Error::from_errno(ret));
+        }
+
+        // SAFETY: The `name` points at a valid c string.
+        let shrinker = unsafe { bindings::shrinker_alloc(0, name.as_char_ptr()) };
+        if shrinker.is_null() {
+            // SAFETY: We initialized it, so its okay to destroy it.
+            unsafe { bindings::list_lru_destroy(self.list_lru.get()) };
+            return Err(Error::from_errno(ret));
+        }
+
+        // SAFETY: We're about to register the shrinker, and these are the fields we need to
+        // initialize. (All other fields are already zeroed.)
+        unsafe {
+            ptr::addr_of_mut!((*shrinker).count_objects).write(Some(rust_shrink_count));
+            ptr::addr_of_mut!((*shrinker).scan_objects).write(Some(rust_shrink_scan));
+        }
+
+        // SAFETY: The new shrinker has been fully initialized, so we can register it.
+        unsafe { bindings::shrinker_register(shrinker) };
+
+        // SAFETY: This initializes the pointer to the shrinker so that we can use it.
+        unsafe { self.inner.get().write(shrinker) };
+
+        Ok(())
+    }
+}
+
+/// A container that manages a page range in a vma.
+///
+/// The pages can be thought of as an array of booleans of whether the pages are usable. The
+/// methods `use_range` and `stop_using_range` set all booleans in a range to true or false
+/// respectively. Initially, no pages are allocated. When a page is not used, it is not freed
+/// immediately. Instead, it is made available to the memory shrinker to free it if the device is
+/// under memory pressure.
+///
+/// It's okay for `use_range` and `stop_using_range` to race with each other, although there's no
+/// way to know whether an index ends up with true or false if a call to `use_range` races with
+/// another call to `stop_using_range` on a given index.
+///
+/// It's also okay for the two methods to race with themselves, e.g. if two threads call
+/// `use_range` on the same index, then that's fine and neither call will return until the page is
+/// allocated and mapped.
+///
+/// The methods that read or write to a range require that the page is marked as in use. So it is
+/// _not_ okay to call `stop_using_range` on a page that is in use by the methods that read or
+/// write to the page.
+#[pin_data(PinnedDrop)]
+pub(crate) struct ShrinkablePageRange {
+    /// Shrinker object registered with the kernel.
+    shrinker: &'static Shrinker,
+    /// Pid using this page range. Only used as debugging information.
+    pid: Pid,
+    /// The mm for the relevant process.
+    mm: ARef<Mm>,
+    /// Used to synchronize calls to `vm_insert_page` and `zap_page_range_single`.
+    #[pin]
+    mm_lock: Mutex<()>,
+    /// Spinlock protecting changes to pages.
+    #[pin]
+    lock: SpinLock<Inner>,
+
+    /// Must not move, since page info has pointers back.
+    #[pin]
+    _pin: PhantomPinned,
+}
+
+struct Inner {
+    /// Array of pages.
+    ///
+    /// Since this is also accessed by the shrinker, we can't use a `Box`, which asserts exclusive
+    /// ownership. To deal with that, we manage it using raw pointers.
+    pages: *mut PageInfo,
+    /// Length of the `pages` array.
+    size: usize,
+    /// The address of the vma to insert the pages into.
+    vma_addr: usize,
+}
+
+unsafe impl Send for ShrinkablePageRange {}
+unsafe impl Sync for ShrinkablePageRange {}
+
+type StableMmGuard =
+    kernel::sync::lock::Guard<'static, (), kernel::sync::lock::mutex::MutexBackend>;
+
+/// An array element that describes the current state of a page.
+///
+/// There are three states:
+///
+///  * Free. The page is None. The `lru` element is not queued.
+///  * Available. The page is Some. The `lru` element is queued to the shrinker's lru.
+///  * Used. The page is Some. The `lru` element is not queued.
+///
+/// When an element is available, the shrinker is able to free the page.
+#[repr(C)]
+struct PageInfo {
+    lru: bindings::list_head,
+    page: Option<Page>,
+    range: *const ShrinkablePageRange,
+}
+
+impl PageInfo {
+    /// # Safety
+    ///
+    /// The caller ensures that reading from `me.page` is ok.
+    unsafe fn has_page(me: *const PageInfo) -> bool {
+        // SAFETY: This pointer offset is in bounds.
+        let page = unsafe { ptr::addr_of!((*me).page) };
+
+        unsafe { (*page).is_some() }
+    }
+
+    /// # Safety
+    ///
+    /// The caller ensures that writing to `me.page` is ok, and that the page is not currently set.
+    unsafe fn set_page(me: *mut PageInfo, page: Page) {
+        // SAFETY: This pointer offset is in bounds.
+        let ptr = unsafe { ptr::addr_of_mut!((*me).page) };
+
+        // SAFETY: The pointer is valid for writing, so also valid for reading.
+        if unsafe { (*ptr).is_some() } {
+            pr_err!("set_page called when there is already a page");
+            // SAFETY: We will initialize the page again below.
+            unsafe { ptr::drop_in_place(ptr) };
+        }
+
+        // SAFETY: The pointer is valid for writing.
+        unsafe { ptr::write(ptr, Some(page)) };
+    }
+
+    /// # Safety
+    ///
+    /// The caller ensures that reading from `me.page` is ok for the duration of 'a.
+    unsafe fn get_page<'a>(me: *const PageInfo) -> Option<&'a Page> {
+        // SAFETY: This pointer offset is in bounds.
+        let ptr = unsafe { ptr::addr_of!((*me).page) };
+
+        // SAFETY: The pointer is valid for reading.
+        unsafe { (*ptr).as_ref() }
+    }
+
+    /// # Safety
+    ///
+    /// The caller ensures that writing to `me.page` is ok for the duration of 'a.
+    unsafe fn take_page(me: *mut PageInfo) -> Option<Page> {
+        // SAFETY: This pointer offset is in bounds.
+        let ptr = unsafe { ptr::addr_of_mut!((*me).page) };
+
+        // SAFETY: The pointer is valid for reading.
+        unsafe { (*ptr).take() }
+    }
+
+    /// Add this page to the lru list, if not already in the list.
+    ///
+    /// # Safety
+    ///
+    /// The pointer must be valid, and it must be the right shrinker.
+    unsafe fn list_lru_add(me: *mut PageInfo, shrinker: &'static Shrinker) {
+        // SAFETY: This pointer offset is in bounds.
+        let lru_ptr = unsafe { ptr::addr_of_mut!((*me).lru) };
+        // SAFETY: The lru pointer is valid, and we're not using it with any other lru list.
+        unsafe { bindings::list_lru_add_obj(shrinker.list_lru.get(), lru_ptr) };
+    }
+
+    /// Remove this page from the lru list, if it is in the list.
+    ///
+    /// # Safety
+    ///
+    /// The pointer must be valid, and it must be the right shrinker.
+    unsafe fn list_lru_del(me: *mut PageInfo, shrinker: &'static Shrinker) {
+        // SAFETY: This pointer offset is in bounds.
+        let lru_ptr = unsafe { ptr::addr_of_mut!((*me).lru) };
+        // SAFETY: The lru pointer is valid, and we're not using it with any other lru list.
+        unsafe { bindings::list_lru_del_obj(shrinker.list_lru.get(), lru_ptr) };
+    }
+}
+
+impl ShrinkablePageRange {
+    /// Create a new `ShrinkablePageRange` using the given shrinker.
+    pub(crate) fn new(shrinker: &'static Shrinker) -> impl PinInit<Self, Error> {
+        try_pin_init!(Self {
+            shrinker,
+            pid: kernel::current!().pid(),
+            mm: ARef::from(&**kernel::current!().mm().ok_or(ESRCH)?),
+            mm_lock <- new_mutex!((), "ShrinkablePageRange::mm"),
+            lock <- new_spinlock!(Inner {
+                pages: ptr::null_mut(),
+                size: 0,
+                vma_addr: 0,
+            }, "ShrinkablePageRange"),
+            _pin: PhantomPinned,
+        })
+    }
+
+    pub(crate) fn stable_trylock_mm(&self) -> Option<StableMmGuard> {
+        // SAFETY: This extends the duration of the reference. Since this call happens before
+        // `mm_lock` is taken in the destructor of `ShrinkablePageRange`, the destructor will block
+        // until the returned guard is dropped. This ensures that the guard is valid until dropped.
+        let mm_lock = unsafe { &*ptr::from_ref(&self.mm_lock) };
+
+        mm_lock.try_lock()
+    }
+
+    /// Register a vma with this page range. Returns the size of the region.
+    pub(crate) fn register_with_vma(&self, vma: &virt::VmAreaNew) -> Result<usize> {
+        let num_bytes = usize::min(vma.end() - vma.start(), bindings::SZ_4M as usize);
+        let num_pages = num_bytes >> PAGE_SHIFT;
+
+        if !ptr::eq::<Mm>(&*self.mm, &**vma.mm()) {
+            pr_debug!("Failed to register with vma: invalid vma->vm_mm");
+            return Err(EINVAL);
+        }
+        if num_pages == 0 {
+            pr_debug!("Failed to register with vma: size zero");
+            return Err(EINVAL);
+        }
+
+        let layout = Layout::array::<PageInfo>(num_pages).map_err(|_| ENOMEM)?;
+        // SAFETY: The layout has non-zero size.
+        let pages = unsafe { alloc::alloc::alloc(layout) as *mut PageInfo };
+        if pages.is_null() {
+            return Err(ENOMEM);
+        }
+
+        // SAFETY: This just initializes the pages array.
+        unsafe {
+            let self_ptr = self as *const ShrinkablePageRange;
+            for i in 0..num_pages {
+                let info = pages.add(i);
+                ptr::addr_of_mut!((*info).range).write(self_ptr);
+                ptr::addr_of_mut!((*info).page).write(None);
+                let lru = ptr::addr_of_mut!((*info).lru);
+                ptr::addr_of_mut!((*lru).next).write(lru);
+                ptr::addr_of_mut!((*lru).prev).write(lru);
+            }
+        }
+
+        let mut inner = self.lock.lock();
+        if inner.size > 0 {
+            pr_debug!("Failed to register with vma: already registered");
+            drop(inner);
+            // SAFETY: The `pages` array was allocated with the same layout.
+            unsafe { alloc::alloc::dealloc(pages.cast(), layout) };
+            return Err(EBUSY);
+        }
+
+        inner.pages = pages;
+        inner.size = num_pages;
+        inner.vma_addr = vma.start();
+
+        Ok(num_pages)
+    }
+
+    /// Make sure that the given pages are allocated and mapped.
+    ///
+    /// Must not be called from an atomic context.
+    pub(crate) fn use_range(&self, start: usize, end: usize) -> Result<()> {
+        crate::trace::trace_update_page_range(self.pid, true, start, end);
+
+        if start >= end {
+            return Ok(());
+        }
+        let mut inner = self.lock.lock();
+        assert!(end <= inner.size);
+
+        for i in start..end {
+            // SAFETY: This pointer offset is in bounds.
+            let page_info = unsafe { inner.pages.add(i) };
+
+            // SAFETY: The pointer is valid, and we hold the lock so reading from the page is okay.
+            if unsafe { PageInfo::has_page(page_info) } {
+                crate::trace::trace_alloc_lru_start(self.pid, i);
+
+                // Since we're going to use the page, we should remove it from the lru list so that
+                // the shrinker will not free it.
+                //
+                // SAFETY: The pointer is valid, and this is the right shrinker.
+                //
+                // The shrinker can't free the page between the check and this call to
+                // `list_lru_del` because we hold the lock.
+                unsafe { PageInfo::list_lru_del(page_info, self.shrinker) };
+
+                crate::trace::trace_alloc_lru_end(self.pid, i);
+            } else {
+                // We have to allocate a new page. Use the slow path.
+                drop(inner);
+                crate::trace::trace_alloc_page_start(self.pid, i);
+                match self.use_page_slow(i) {
+                    Ok(()) => {}
+                    Err(err) => {
+                        pr_warn!("Error in use_page_slow: {:?}", err);
+                        return Err(err);
+                    }
+                }
+                crate::trace::trace_alloc_page_end(self.pid, i);
+                inner = self.lock.lock();
+            }
+        }
+        Ok(())
+    }
+
+    /// Mark the given page as in use, slow path.
+    ///
+    /// Must not be called from an atomic context.
+    ///
+    /// # Safety
+    ///
+    /// Assumes that `i` is in bounds.
+    #[cold]
+    fn use_page_slow(&self, i: usize) -> Result<()> {
+        let new_page = Page::alloc_page(GFP_KERNEL | __GFP_HIGHMEM | __GFP_ZERO)?;
+
+        let mm_mutex = self.mm_lock.lock();
+        let inner = self.lock.lock();
+
+        // SAFETY: This pointer offset is in bounds.
+        let page_info = unsafe { inner.pages.add(i) };
+
+        // SAFETY: The pointer is valid, and we hold the lock so reading from the page is okay.
+        if unsafe { PageInfo::has_page(page_info) } {
+            // The page was already there, or someone else added the page while we didn't hold the
+            // spinlock.
+            //
+            // SAFETY: The pointer is valid, and this is the right shrinker.
+            //
+            // The shrinker can't free the page between the check and this call to
+            // `list_lru_del` because we hold the lock.
+            unsafe { PageInfo::list_lru_del(page_info, self.shrinker) };
+            return Ok(());
+        }
+
+        let vma_addr = inner.vma_addr;
+        // Release the spinlock while we insert the page into the vma.
+        drop(inner);
+
+        // No overflow since we stay in bounds of the vma.
+        let user_page_addr = vma_addr + (i << PAGE_SHIFT);
+
+        // We use `mmput_async` when dropping the `mm` because `use_page_slow` is usually used from
+        // a remote process. If the call to `mmput` races with the process shutting down, then the
+        // caller of `use_page_slow` becomes responsible for cleaning up the `mm`, which doesn't
+        // happen until it returns to userspace. However, the caller might instead go to sleep and
+        // wait for the owner of the `mm` to wake it up, which doesn't happen because it's in the
+        // middle of a shutdown process that wont complete until the `mm` is dropped. This can
+        // amount to a deadlock.
+        //
+        // Using `mmput_async` avoids this, because then the `mm` cleanup is instead queued to a
+        // workqueue.
+        MmWithUser::into_mmput_async(self.mm.mmget_not_zero().ok_or(ESRCH)?)
+            .mmap_read_lock()
+            .vma_lookup(vma_addr)
+            .ok_or(ESRCH)?
+            .as_mixedmap_vma()
+            .ok_or(ESRCH)?
+            .vm_insert_page(user_page_addr, &new_page)
+            .inspect_err(|err| {
+                pr_warn!(
+                    "Failed to vm_insert_page({}): vma_addr:{} i:{} err:{:?}",
+                    user_page_addr,
+                    vma_addr,
+                    i,
+                    err
+                )
+            })?;
+
+        let inner = self.lock.lock();
+
+        // SAFETY: The `page_info` pointer is valid and currently does not have a page. The page
+        // can be written to since we hold the lock.
+        //
+        // We released and reacquired the spinlock since we checked that the page is null, but we
+        // always hold the mm_lock mutex when setting the page to a non-null value, so it's not
+        // possible for someone else to have changed it since our check.
+        unsafe { PageInfo::set_page(page_info, new_page) };
+
+        drop(inner);
+        drop(mm_mutex);
+
+        Ok(())
+    }
+
+    /// If the given page is in use, then mark it as available so that the shrinker can free it.
+    ///
+    /// May be called from an atomic context.
+    pub(crate) fn stop_using_range(&self, start: usize, end: usize) {
+        crate::trace::trace_update_page_range(self.pid, false, start, end);
+
+        if start >= end {
+            return;
+        }
+        let inner = self.lock.lock();
+        assert!(end <= inner.size);
+
+        for i in (start..end).rev() {
+            // SAFETY: The pointer is in bounds.
+            let page_info = unsafe { inner.pages.add(i) };
+
+            // SAFETY: Okay for reading since we have the lock.
+            if unsafe { PageInfo::has_page(page_info) } {
+                crate::trace::trace_free_lru_start(self.pid, i);
+
+                // SAFETY: The pointer is valid, and it's the right shrinker.
+                unsafe { PageInfo::list_lru_add(page_info, self.shrinker) };
+
+                crate::trace::trace_free_lru_end(self.pid, i);
+            }
+        }
+    }
+
+    /// Helper for reading or writing to a range of bytes that may overlap with several pages.
+    ///
+    /// # Safety
+    ///
+    /// All pages touched by this operation must be in use for the duration of this call.
+    unsafe fn iterate<T>(&self, mut offset: usize, mut size: usize, mut cb: T) -> Result
+    where
+        T: FnMut(&Page, usize, usize) -> Result,
+    {
+        if size == 0 {
+            return Ok(());
+        }
+
+        // SAFETY: The caller promises that the pages touched by this call are in use. It's only
+        // possible for a page to be in use if we have already been registered with a vma, and we
+        // only change the `pages` and `size` fields during registration with a vma, so there is no
+        // race when we read them here without taking the lock.
+        let (pages, num_pages) = {
+            let inner = self.lock.lock();
+            (inner.pages, inner.size)
+        };
+        let num_bytes = num_pages << PAGE_SHIFT;
+
+        // Check that the request is within the buffer.
+        if offset.checked_add(size).ok_or(EFAULT)? > num_bytes {
+            return Err(EFAULT);
+        }
+
+        let mut page_index = offset >> PAGE_SHIFT;
+        offset &= PAGE_SIZE - 1;
+        while size > 0 {
+            let available = usize::min(size, PAGE_SIZE - offset);
+            // SAFETY: The pointer is in bounds.
+            let page_info = unsafe { pages.add(page_index) };
+            // SAFETY: The caller guarantees that this page is in the "in use" state for the
+            // duration of this call to `iterate`, so nobody will change the page.
+            let page = unsafe { PageInfo::get_page(page_info) };
+            if page.is_none() {
+                pr_warn!("Page is null!");
+            }
+            let page = page.ok_or(EFAULT)?;
+            cb(page, offset, available)?;
+            size -= available;
+            page_index += 1;
+            offset = 0;
+        }
+        Ok(())
+    }
+
+    /// Copy from userspace into this page range.
+    ///
+    /// # Safety
+    ///
+    /// All pages touched by this operation must be in use for the duration of this call.
+    pub(crate) unsafe fn copy_from_user_slice(
+        &self,
+        reader: &mut UserSliceReader,
+        offset: usize,
+        size: usize,
+    ) -> Result {
+        // SAFETY: `self.iterate` has the same safety requirements as `copy_from_user_slice`.
+        unsafe {
+            self.iterate(offset, size, |page, offset, to_copy| {
+                page.copy_from_user_slice_raw(reader, offset, to_copy)
+            })
+        }
+    }
+
+    /// Copy from this page range into kernel space.
+    ///
+    /// # Safety
+    ///
+    /// All pages touched by this operation must be in use for the duration of this call.
+    pub(crate) unsafe fn read<T: FromBytes>(&self, offset: usize) -> Result<T> {
+        let mut out = MaybeUninit::<T>::uninit();
+        let mut out_offset = 0;
+        // SAFETY: `self.iterate` has the same safety requirements as `read`.
+        unsafe {
+            self.iterate(offset, size_of::<T>(), |page, offset, to_copy| {
+                // SAFETY: The sum of `offset` and `to_copy` is bounded by the size of T.
+                let obj_ptr = (out.as_mut_ptr() as *mut u8).add(out_offset);
+                // SAFETY: The pointer points is in-bounds of the `out` variable, so it is valid.
+                page.read_raw(obj_ptr, offset, to_copy)?;
+                out_offset += to_copy;
+                Ok(())
+            })?;
+        }
+        // SAFETY: We just initialised the data.
+        Ok(unsafe { out.assume_init() })
+    }
+
+    /// Copy from kernel space into this page range.
+    ///
+    /// # Safety
+    ///
+    /// All pages touched by this operation must be in use for the duration of this call.
+    pub(crate) unsafe fn write<T: ?Sized>(&self, offset: usize, obj: &T) -> Result {
+        let mut obj_offset = 0;
+        // SAFETY: `self.iterate` has the same safety requirements as `write`.
+        unsafe {
+            self.iterate(offset, size_of_val(obj), |page, offset, to_copy| {
+                // SAFETY: The sum of `offset` and `to_copy` is bounded by the size of T.
+                let obj_ptr = (obj as *const T as *const u8).add(obj_offset);
+                // SAFETY: We have a reference to the object, so the pointer is valid.
+                page.write_raw(obj_ptr, offset, to_copy)?;
+                obj_offset += to_copy;
+                Ok(())
+            })
+        }
+    }
+
+    /// Write zeroes to the given range.
+    ///
+    /// # Safety
+    ///
+    /// All pages touched by this operation must be in use for the duration of this call.
+    pub(crate) unsafe fn fill_zero(&self, offset: usize, size: usize) -> Result {
+        // SAFETY: `self.iterate` has the same safety requirements as `copy_into`.
+        unsafe {
+            self.iterate(offset, size, |page, offset, len| {
+                page.fill_zero_raw(offset, len)
+            })
+        }
+    }
+}
+
+#[pinned_drop]
+impl PinnedDrop for ShrinkablePageRange {
+    fn drop(self: Pin<&mut Self>) {
+        let (pages, size) = {
+            let lock = self.lock.lock();
+            (lock.pages, lock.size)
+        };
+
+        if size == 0 {
+            return;
+        }
+
+        // This is the destructor, so unlike the other methods, we only need to worry about races
+        // with the shrinker here.
+        for i in 0..size {
+            // SAFETY: The pointer is valid and it's the right shrinker.
+            unsafe { PageInfo::list_lru_del(pages.add(i), self.shrinker) };
+            // SAFETY: If the shrinker was going to free this page, then it would have taken it
+            // from the PageInfo before releasing the lru lock. Thus, the call to `list_lru_del`
+            // will either remove it before the shrinker can access it, or the shrinker will
+            // already have taken the page at this point.
+            unsafe { drop(PageInfo::take_page(pages.add(i))) };
+        }
+
+        // Wait for users of the mutex to go away. This call is necessary for the safety of
+        // `stable_trylock_mm`.
+        drop(self.mm_lock.lock());
+
+        // SAFETY: This computation did not overflow when allocating the pages array, so it will
+        // not overflow this time.
+        let layout = unsafe { Layout::array::<PageInfo>(size).unwrap_unchecked() };
+
+        // SAFETY: The `pages` array was allocated with the same layout.
+        unsafe { alloc::alloc::dealloc(pages.cast(), layout) };
+    }
+}
+
+#[no_mangle]
+unsafe extern "C" fn rust_shrink_count(
+    shrink: *mut bindings::shrinker,
+    _sc: *mut bindings::shrink_control,
+) -> c_ulong {
+    // SAFETY: This method is only used with the `Shrinker` type, and the cast is valid since
+    // `shrinker` is the first field of a #[repr(C)] struct.
+    let shrinker = unsafe { &*shrink.cast::<Shrinker>() };
+    // SAFETY: Accessing the lru list is okay. Just an FFI call.
+    unsafe { bindings::list_lru_count(shrinker.list_lru.get()) }
+}
+
+#[no_mangle]
+unsafe extern "C" fn rust_shrink_scan(
+    shrink: *mut bindings::shrinker,
+    sc: *mut bindings::shrink_control,
+) -> c_ulong {
+    // SAFETY: This method is only used with the `Shrinker` type, and the cast is valid since
+    // `shrinker` is the first field of a #[repr(C)] struct.
+    let shrinker = unsafe { &*shrink.cast::<Shrinker>() };
+    // SAFETY: Caller guarantees that it is safe to read this field.
+    let nr_to_scan = unsafe { (*sc).nr_to_scan };
+    // SAFETY: Accessing the lru list is okay. Just an FFI call.
+    unsafe {
+        extern "C" {
+            fn rust_shrink_free_page_wrap(
+                item: *mut bindings::list_head,
+                list: *mut bindings::list_lru_one,
+                lock: *mut bindings::spinlock_t,
+                cb_arg: *mut core::ffi::c_void,
+            ) -> bindings::lru_status;
+        }
+
+        bindings::list_lru_walk(
+            shrinker.list_lru.get(),
+            Some(rust_shrink_free_page_wrap),
+            ptr::null_mut(),
+            nr_to_scan,
+        )
+    }
+}
+
+const LRU_SKIP: bindings::lru_status = bindings::lru_status_LRU_SKIP;
+const LRU_REMOVED_ENTRY: bindings::lru_status = bindings::lru_status_LRU_REMOVED_RETRY;
+
+#[no_mangle]
+unsafe extern "C" fn rust_shrink_free_page(
+    item: *mut bindings::list_head,
+    lru: *mut bindings::list_lru_one,
+    lru_lock: *mut bindings::spinlock_t,
+    _cb_arg: *mut c_void,
+) -> bindings::lru_status {
+    // Fields that should survive after unlocking the lru lock.
+    let pid;
+    let page;
+    let page_index;
+    let mm;
+    let mmap_read;
+    let mm_mutex;
+    let vma_addr;
+
+    {
+        // SAFETY: The `list_head` field is first in `PageInfo`.
+        let info = item as *mut PageInfo;
+        let range = unsafe { &*((*info).range) };
+
+        mm = match range.mm.mmget_not_zero() {
+            Some(mm) => MmWithUser::into_mmput_async(mm),
+            None => return LRU_SKIP,
+        };
+
+        mm_mutex = match range.stable_trylock_mm() {
+            Some(guard) => guard,
+            None => return LRU_SKIP,
+        };
+
+        mmap_read = match mm.mmap_read_trylock() {
+            Some(guard) => guard,
+            None => return LRU_SKIP,
+        };
+
+        // We can't lock it normally here, since we hold the lru lock.
+        let inner = match range.lock.try_lock() {
+            Some(inner) => inner,
+            None => return LRU_SKIP,
+        };
+
+        // SAFETY: The item is in this lru list, so it's okay to remove it.
+        unsafe { bindings::list_lru_isolate(lru, item) };
+
+        // SAFETY: Both pointers are in bounds of the same allocation.
+        page_index = unsafe { info.offset_from(inner.pages) } as usize;
+        pid = range.pid;
+
+        crate::trace::trace_unmap_kernel_start(pid, page_index);
+
+        // SAFETY: We hold the spinlock, so we can take the page.
+        //
+        // This sets the page pointer to zero before we unmap it from the vma. However, we call
+        // `zap_page_range` before we release the mmap lock, so `use_page_slow` will not be able to
+        // insert a new page until after our call to `zap_page_range`.
+        page = unsafe { PageInfo::take_page(info) };
+        vma_addr = inner.vma_addr;
+
+        crate::trace::trace_unmap_kernel_end(pid, page_index);
+
+        // From this point on, we don't access this PageInfo or ShrinkablePageRange again, because
+        // they can be freed at any point after we unlock `lru_lock`. This is with the exception of
+        // `mm_mutex` which is kept alive by holding the lock.
+    }
+
+    // SAFETY: The lru lock is locked when this method is called.
+    unsafe { bindings::spin_unlock(lru_lock) };
+
+    if let Some(vma) = mmap_read.vma_lookup(vma_addr) {
+        let user_page_addr = vma_addr + (page_index << PAGE_SHIFT);
+        crate::trace::trace_unmap_user_start(pid, page_index);
+        vma.zap_page_range_single(user_page_addr, PAGE_SIZE);
+        crate::trace::trace_unmap_user_end(pid, page_index);
+    }
+
+    drop(mmap_read);
+    drop(mm_mutex);
+    drop(mm);
+    drop(page);
+
+    // SAFETY: We just unlocked the lru lock, but it should be locked when we return.
+    unsafe { bindings::spin_lock(lru_lock) };
+
+    LRU_REMOVED_ENTRY
+}

drivers/android/binder/page_range_helper.c

@@ -0,0 +1,25 @@
+// SPDX-License-Identifier: GPL-2.0
+
+/* C helper for page_range.rs to work around a CFI violation.
+ *
+ * Bindgen currently pretends that `enum lru_status` is the same as an integer.
+ * This assumption is fine ABI-wise, but once you add CFI to the mix, it
+ * triggers a CFI violation because `enum lru_status` gets a different CFI tag.
+ *
+ * This file contains a workaround until bindgen can be fixed.
+ *
+ * Copyright (C) 2024 Google LLC.
+ */
+#include <linux/list_lru.h>
+#include <linux/spinlock.h>
+
+unsigned int rust_shrink_free_page(struct list_head *item,
+				   struct list_lru_one *list, spinlock_t *lock,
+				   void *cb_arg);
+
+enum lru_status
+rust_shrink_free_page_wrap(struct list_head *item, struct list_lru_one *list,
+			   spinlock_t *lock, void *cb_arg)
+{
+	return rust_shrink_free_page(item, list, lock, cb_arg);
+}

drivers/android/binder/prio.rs

@@ -0,0 +1,80 @@
+// SPDX-License-Identifier: GPL-2.0
+
+// Copyright (C) 2024 Google LLC.
+
+//! This module defines the types and methods relevant to priority inheritance.
+
+use kernel::bindings;
+
+pub(crate) type Policy = core::ffi::c_uint;
+pub(crate) type Priority = core::ffi::c_int;
+pub(crate) type Nice = core::ffi::c_int;
+
+pub(crate) const SCHED_NORMAL: Policy = bindings::SCHED_NORMAL;
+pub(crate) const SCHED_FIFO: Policy = bindings::SCHED_FIFO;
+pub(crate) const MIN_NICE: Nice = bindings::MIN_NICE as _;
+pub(crate) const MAX_NICE: Nice = bindings::MAX_NICE as _;
+pub(crate) const DEFAULT_PRIO: Priority = bindings::DEFAULT_PRIO as _;
+pub(crate) const MAX_RT_PRIO: Priority = bindings::MAX_RT_PRIO as _;
+
+/// Scheduler policy and priority.
+///
+/// The binder driver supports inheriting the following scheduler policies:
+/// * SCHED_NORMAL
+/// * SCHED_BATCH
+/// * SCHED_FIFO
+/// * SCHED_RR
+#[derive(Copy, Clone, Default)]
+pub(crate) struct BinderPriority {
+    pub(crate) sched_policy: Policy,
+    pub(crate) prio: Priority,
+}
+
+#[derive(Copy, Clone, Eq, PartialEq)]
+pub(crate) enum PriorityState {
+    Set,
+    Pending,
+    Abort,
+}
+
+pub(crate) fn get_default_prio_from_task(task: &kernel::task::Task) -> BinderPriority {
+    if is_supported_policy(task.policy()) {
+        BinderPriority {
+            sched_policy: task.policy(),
+            prio: task.normal_prio(),
+        }
+    } else {
+        BinderPriority {
+            sched_policy: SCHED_NORMAL,
+            prio: DEFAULT_PRIO,
+        }
+    }
+}
+
+pub(crate) fn is_rt_policy(policy: Policy) -> bool {
+    policy == bindings::SCHED_FIFO || policy == bindings::SCHED_RR
+}
+
+pub(crate) fn is_fair_policy(policy: Policy) -> bool {
+    policy == bindings::SCHED_NORMAL || policy == bindings::SCHED_BATCH
+}
+
+pub(crate) fn is_supported_policy(policy: Policy) -> bool {
+    is_fair_policy(policy) || is_rt_policy(policy)
+}
+
+pub(crate) fn to_userspace_prio(policy: Policy, prio: Priority) -> Nice {
+    if is_fair_policy(policy) {
+        prio - DEFAULT_PRIO
+    } else {
+        MAX_RT_PRIO - 1 - prio
+    }
+}
+
+pub(crate) fn to_kernel_prio(policy: Policy, prio: Nice) -> Priority {
+    if is_fair_policy(policy) {
+        prio + DEFAULT_PRIO
+    } else {
+        MAX_RT_PRIO - 1 - prio
+    }
+}

drivers/android/binder/range_alloc/array.rs

@@ -0,0 +1,282 @@
+// SPDX-License-Identifier: GPL-2.0
+
+// Copyright (C) 2024 Google LLC.
+
+use kernel::{
+    page::{PAGE_MASK, PAGE_SIZE},
+    prelude::*,
+    seq_file::SeqFile,
+    seq_print,
+    task::Pid,
+};
+
+use crate::range_alloc::{DescriptorState, FreedRange, Range};
+
+/// Keeps track of allocations in a process' mmap.
+///
+/// Each process has an mmap where the data for incoming transactions will be placed. This struct
+/// keeps track of allocations made in the mmap. For each allocation, we store a descriptor that
+/// has metadata related to the allocation. We also keep track of available free space.
+pub(super) struct ArrayRangeAllocator<T> {
+    /// This stores all ranges that are allocated. Unlike the tree based allocator, we do *not*
+    /// store the free ranges.
+    ///
+    /// Sorted by offset.
+    pub(super) ranges: Vec<Range<T>>,
+    size: usize,
+    free_oneway_space: usize,
+}
+
+struct FindEmptyRes {
+    /// Which index in `ranges` should we insert the new range at?
+    ///
+    /// Inserting the new range at this index keeps `ranges` sorted.
+    insert_at_idx: usize,
+    /// Which offset should we insert the new range at?
+    insert_at_offset: usize,
+}
+
+impl<T> ArrayRangeAllocator<T> {
+    pub(crate) fn new(size: usize, alloc: EmptyArrayAlloc<T>) -> Self {
+        Self {
+            ranges: alloc.ranges,
+            size,
+            free_oneway_space: size / 2,
+        }
+    }
+
+    pub(crate) fn free_oneway_space(&self) -> usize {
+        self.free_oneway_space
+    }
+
+    pub(crate) fn count_buffers(&self) -> usize {
+        self.ranges.len()
+    }
+
+    pub(crate) fn total_size(&self) -> usize {
+        self.size
+    }
+
+    pub(crate) fn is_full(&self) -> bool {
+        self.ranges.len() == self.ranges.capacity()
+    }
+
+    pub(crate) fn debug_print(&self, m: &SeqFile) -> Result<()> {
+        for range in &self.ranges {
+            seq_print!(
+                m,
+                "  buffer {}: {} size {} pid {} oneway {}",
+                0,
+                range.offset,
+                range.size,
+                range.state.pid(),
+                range.state.is_oneway(),
+            );
+            if let DescriptorState::Reserved(_) = range.state {
+                seq_print!(m, " reserved\n");
+            } else {
+                seq_print!(m, " allocated\n");
+            }
+        }
+        Ok(())
+    }
+
+    /// Find somewhere to put a new range.
+    ///
+    /// Unlike the tree implementation, we do not bother to find the smallest gap. The idea is that
+    /// fragmentation isn't a big issue when we don't have many ranges.
+    ///
+    /// Returns the index that the new range should have in `self.ranges` after insertion.
+    fn find_empty_range(&self, size: usize) -> Option<FindEmptyRes> {
+        let after_last_range = self.ranges.last().map(Range::endpoint).unwrap_or(0);
+
+        if size <= self.total_size() - after_last_range {
+            // We can put the range at the end, so just do that.
+            Some(FindEmptyRes {
+                insert_at_idx: self.ranges.len(),
+                insert_at_offset: after_last_range,
+            })
+        } else {
+            let mut end_of_prev = 0;
+            for (i, range) in self.ranges.iter().enumerate() {
+                // Does it fit before the i'th range?
+                if size <= range.offset - end_of_prev {
+                    return Some(FindEmptyRes {
+                        insert_at_idx: i,
+                        insert_at_offset: end_of_prev,
+                    });
+                }
+                end_of_prev = range.endpoint();
+            }
+            None
+        }
+    }
+
+    pub(crate) fn reserve_new(
+        &mut self,
+        debug_id: usize,
+        size: usize,
+        is_oneway: bool,
+        pid: Pid,
+    ) -> Result<usize> {
+        // Compute new value of free_oneway_space, which is set only on success.
+        let new_oneway_space = if is_oneway {
+            match self.free_oneway_space.checked_sub(size) {
+                Some(new_oneway_space) => new_oneway_space,
+                None => return Err(ENOSPC),
+            }
+        } else {
+            self.free_oneway_space
+        };
+
+        let FindEmptyRes {
+            insert_at_idx,
+            insert_at_offset,
+        } = self.find_empty_range(size).ok_or(ENOSPC)?;
+        self.free_oneway_space = new_oneway_space;
+
+        let new_range = Range {
+            offset: insert_at_offset,
+            size,
+            state: DescriptorState::new(is_oneway, debug_id, pid),
+        };
+        // Insert the value at the given index to keep the array sorted.
+        insert_within_capacity(&mut self.ranges, insert_at_idx, new_range);
+
+        Ok(insert_at_offset)
+    }
+
+    pub(crate) fn reservation_abort(&mut self, offset: usize) -> Result<FreedRange> {
+        // This could use a binary search, but linear scans are usually faster for small arrays.
+        let i = self
+            .ranges
+            .iter()
+            .position(|range| range.offset == offset)
+            .ok_or(EINVAL)?;
+        let range = &self.ranges[i];
+
+        if let DescriptorState::Allocated(_) = range.state {
+            return Err(EPERM);
+        }
+
+        let size = range.size;
+        let offset = range.offset;
+
+        if range.state.is_oneway() {
+            self.free_oneway_space += size;
+        }
+
+        // This computes the range of pages that are no longer used by *any* allocated range. The
+        // caller will mark them as unused, which means that they can be freed if the system comes
+        // under memory pressure.
+        let mut freed_range = FreedRange::interior_pages(offset, size);
+        if offset % PAGE_SIZE != 0 {
+            if i == 0 || self.ranges[i - 1].endpoint() <= (offset & PAGE_MASK) {
+                freed_range.start_page_idx -= 1;
+            }
+        }
+        if range.endpoint() % PAGE_SIZE != 0 {
+            let page_after = (range.endpoint() & PAGE_MASK) + PAGE_SIZE;
+            if i + 1 == self.ranges.len() || page_after <= self.ranges[i + 1].offset {
+                freed_range.end_page_idx += 1;
+            }
+        }
+
+        self.ranges.remove(i);
+        Ok(freed_range)
+    }
+
+    pub(crate) fn reservation_commit(&mut self, offset: usize, data: Option<T>) -> Result {
+        // This could use a binary search, but linear scans are usually faster for small arrays.
+        let range = self
+            .ranges
+            .iter_mut()
+            .find(|range| range.offset == offset)
+            .ok_or(ENOENT)?;
+
+        let DescriptorState::Reserved(reservation) = &range.state else {
+            return Err(ENOENT);
+        };
+
+        range.state = DescriptorState::Allocated(reservation.clone().allocate(data));
+        Ok(())
+    }
+
+    pub(crate) fn reserve_existing(&mut self, offset: usize) -> Result<(usize, usize, Option<T>)> {
+        // This could use a binary search, but linear scans are usually faster for small arrays.
+        let range = self
+            .ranges
+            .iter_mut()
+            .find(|range| range.offset == offset)
+            .ok_or(ENOENT)?;
+
+        let DescriptorState::Allocated(allocation) = &mut range.state else {
+            return Err(ENOENT);
+        };
+
+        let data = allocation.take();
+        let debug_id = allocation.reservation.debug_id;
+        range.state = DescriptorState::Reserved(allocation.reservation.clone());
+        Ok((range.size, debug_id, data))
+    }
+
+    pub(crate) fn take_for_each<F: Fn(usize, usize, usize, Option<T>)>(&mut self, callback: F) {
+        for range in self.ranges.iter_mut() {
+            if let DescriptorState::Allocated(allocation) = &mut range.state {
+                callback(
+                    range.offset,
+                    range.size,
+                    allocation.reservation.debug_id,
+                    allocation.data.take(),
+                );
+            }
+        }
+    }
+}
+
+pub(crate) struct EmptyArrayAlloc<T> {
+    ranges: Vec<Range<T>>,
+}
+
+impl<T> EmptyArrayAlloc<T> {
+    pub(crate) fn try_new(capacity: usize) -> Result<Self> {
+        Ok(Self {
+            ranges: Vec::with_capacity(capacity, GFP_KERNEL)?,
+        })
+    }
+}
+
+/// Copied from `rust/alloc/vec/mod.rs` with allocation codepath removed.
+///
+/// TODO: Either add this to the standard library (like [`push_within_capacity`]) or move it to the
+/// kernel crate once [the updated allocation APIs][alloc] are available.
+///
+/// [`push_within_capacity`]: https://github.com/rust-lang/rust/issues/100486
+/// [alloc]: https://lore.kernel.org/r/20240328013603.206764-1-wedsonaf@gmail.com
+fn insert_within_capacity<T>(vec: &mut Vec<T>, index: usize, element: T) {
+    let len = vec.len();
+
+    if len == vec.capacity() {
+        panic!("no capacity to insert");
+    }
+
+    unsafe {
+        // The spot to put the new value
+        {
+            let p = vec.as_mut_ptr().add(index);
+            if index < len {
+                // Shift everything over to make space. (Duplicating the
+                // `index`th element into two consecutive places.)
+                core::ptr::copy(p, p.add(1), len - index);
+            } else if index == len {
+                // No elements need shifting.
+            } else {
+                panic!("insertion index (is {index}) should be <= len (is {len})");
+            }
+            // Write it in, overwriting the first copy of the `index`th
+            // element.
+            core::ptr::write(p, element);
+        }
+        vec.set_len(len + 1);
+    }
+}

drivers/android/binder/range_alloc/mod.rs

@@ -0,0 +1,326 @@
+// SPDX-License-Identifier: GPL-2.0
+
+// Copyright (C) 2024 Google LLC.
+
+use kernel::{page::PAGE_SIZE, prelude::*, seq_file::SeqFile, task::Pid};
+
+mod tree;
+use self::tree::{FromArrayAllocs, ReserveNewTreeAlloc, TreeRangeAllocator};
+
+mod array;
+use self::array::{ArrayRangeAllocator, EmptyArrayAlloc};
+
+enum DescriptorState<T> {
+    Reserved(Reservation),
+    Allocated(Allocation<T>),
+}
+
+impl<T> DescriptorState<T> {
+    fn new(is_oneway: bool, debug_id: usize, pid: Pid) -> Self {
+        DescriptorState::Reserved(Reservation {
+            debug_id,
+            is_oneway,
+            pid,
+        })
+    }
+
+    fn pid(&self) -> Pid {
+        match self {
+            DescriptorState::Reserved(inner) => inner.pid,
+            DescriptorState::Allocated(inner) => inner.reservation.pid,
+        }
+    }
+
+    fn is_oneway(&self) -> bool {
+        match self {
+            DescriptorState::Reserved(inner) => inner.is_oneway,
+            DescriptorState::Allocated(inner) => inner.reservation.is_oneway,
+        }
+    }
+}
+
+#[derive(Clone)]
+struct Reservation {
+    debug_id: usize,
+    is_oneway: bool,
+    pid: Pid,
+}
+
+impl Reservation {
+    fn allocate<T>(self, data: Option<T>) -> Allocation<T> {
+        Allocation {
+            data,
+            reservation: self,
+        }
+    }
+}
+
+struct Allocation<T> {
+    reservation: Reservation,
+    data: Option<T>,
+}
+
+impl<T> Allocation<T> {
+    fn deallocate(self) -> (Reservation, Option<T>) {
+        (self.reservation, self.data)
+    }
+
+    fn debug_id(&self) -> usize {
+        self.reservation.debug_id
+    }
+
+    fn take(&mut self) -> Option<T> {
+        self.data.take()
+    }
+}
+
+/// The array implementation must switch to the tree if it wants to go beyond this number of
+/// ranges.
+const TREE_THRESHOLD: usize = 8;
+
+/// Represents a range of pages that have just become completely free.
+#[derive(Copy, Clone)]
+pub(crate) struct FreedRange {
+    pub(crate) start_page_idx: usize,
+    pub(crate) end_page_idx: usize,
+}
+
+impl FreedRange {
+    fn interior_pages(offset: usize, size: usize) -> FreedRange {
+        FreedRange {
+            // Divide round up
+            start_page_idx: (offset + (PAGE_SIZE - 1)) / PAGE_SIZE,
+            // Divide round down
+            end_page_idx: (offset + size) / PAGE_SIZE,
+        }
+    }
+}
+
+struct Range<T> {
+    offset: usize,
+    size: usize,
+    state: DescriptorState<T>,
+}
+
+impl<T> Range<T> {
+    fn endpoint(&self) -> usize {
+        self.offset + self.size
+    }
+}
+
+pub(crate) struct RangeAllocator<T> {
+    inner: Impl<T>,
+}
+
+enum Impl<T> {
+    Empty(usize),
+    Array(ArrayRangeAllocator<T>),
+    Tree(TreeRangeAllocator<T>),
+}
+
+impl<T> RangeAllocator<T> {
+    pub(crate) fn new(size: usize) -> Self {
+        Self {
+            inner: Impl::Empty(size),
+        }
+    }
+
+    pub(crate) fn free_oneway_space(&self) -> usize {
+        match &self.inner {
+            Impl::Empty(size) => size / 2,
+            Impl::Array(array) => array.free_oneway_space(),
+            Impl::Tree(tree) => tree.free_oneway_space(),
+        }
+    }
+
+    pub(crate) fn count_buffers(&self) -> usize {
+        match &self.inner {
+            Impl::Empty(_size) => 0,
+            Impl::Array(array) => array.count_buffers(),
+            Impl::Tree(tree) => tree.count_buffers(),
+        }
+    }
+
+    pub(crate) fn debug_print(&self, m: &SeqFile) -> Result<()> {
+        match &self.inner {
+            Impl::Empty(_size) => Ok(()),
+            Impl::Array(array) => array.debug_print(m),
+            Impl::Tree(tree) => tree.debug_print(m),
+        }
+    }
+
+    /// Try to reserve a new buffer, using the provided allocation if necessary.
+    pub(crate) fn reserve_new(&mut self, mut args: ReserveNewArgs<T>) -> Result<ReserveNew<T>> {
+        match &mut self.inner {
+            Impl::Empty(size) => {
+                let empty_array = match args.empty_array_alloc.take() {
+                    Some(empty_array) => ArrayRangeAllocator::new(*size, empty_array),
+                    None => {
+                        return Ok(ReserveNew::NeedAlloc(ReserveNewNeedAlloc {
+                            args,
+                            need_empty_array_alloc: true,
+                            need_new_tree_alloc: false,
+                            need_tree_alloc: false,
+                        }))
+                    }
+                };
+
+                self.inner = Impl::Array(empty_array);
+                self.reserve_new(args)
+            }
+            Impl::Array(array) if array.is_full() => {
+                let allocs = match args.new_tree_alloc {
+                    Some(ref mut allocs) => allocs,
+                    None => {
+                        return Ok(ReserveNew::NeedAlloc(ReserveNewNeedAlloc {
+                            args,
+                            need_empty_array_alloc: false,
+                            need_new_tree_alloc: true,
+                            need_tree_alloc: true,
+                        }))
+                    }
+                };
+
+                let new_tree =
+                    TreeRangeAllocator::from_array(array.total_size(), &mut array.ranges, allocs);
+
+                self.inner = Impl::Tree(new_tree);
+                self.reserve_new(args)
+            }
+            Impl::Array(array) => {
+                let offset =
+                    array.reserve_new(args.debug_id, args.size, args.is_oneway, args.pid)?;
+                Ok(ReserveNew::Success(ReserveNewSuccess {
+                    offset,
+                    oneway_spam_detected: false,
+                    _empty_array_alloc: args.empty_array_alloc,
+                    _new_tree_alloc: args.new_tree_alloc,
+                    _tree_alloc: args.tree_alloc,
+                }))
+            }
+            Impl::Tree(tree) => {
+                let alloc = match args.tree_alloc {
+                    Some(alloc) => alloc,
+                    None => {
+                        return Ok(ReserveNew::NeedAlloc(ReserveNewNeedAlloc {
+                            args,
+                            need_empty_array_alloc: false,
+                            need_new_tree_alloc: false,
+                            need_tree_alloc: true,
+                        }));
+                    }
+                };
+                let (offset, oneway_spam_detected) =
+                    tree.reserve_new(args.debug_id, args.size, args.is_oneway, args.pid, alloc)?;
+                Ok(ReserveNew::Success(ReserveNewSuccess {
+                    offset,
+                    oneway_spam_detected,
+                    _empty_array_alloc: args.empty_array_alloc,
+                    _new_tree_alloc: args.new_tree_alloc,
+                    _tree_alloc: None,
+                }))
+            }
+        }
+    }
+
+    /// Deletes the allocations at `offset`.
+    pub(crate) fn reservation_abort(&mut self, offset: usize) -> Result<FreedRange> {
+        match &mut self.inner {
+            Impl::Empty(_size) => Err(EINVAL),
+            Impl::Array(array) => array.reservation_abort(offset),
+            Impl::Tree(tree) => {
+                let freed_range = tree.reservation_abort(offset)?;
+                if tree.is_empty() {
+                    self.inner = Impl::Empty(tree.total_size());
+                }
+                Ok(freed_range)
+            }
+        }
+    }
+
+    /// Called when an allocation is no longer in use by the kernel.
+    pub(crate) fn reservation_commit(&mut self, offset: usize, data: Option<T>) -> Result {
+        match &mut self.inner {
+            Impl::Empty(_size) => Err(EINVAL),
+            Impl::Array(array) => array.reservation_commit(offset, data),
+            Impl::Tree(tree) => tree.reservation_commit(offset, data),
+        }
+    }
+
+    /// Called when the kernel starts using an allocation.
+    ///
+    /// Returns the size of the existing entry and the data associated with it.
+    pub(crate) fn reserve_existing(&mut self, offset: usize) -> Result<(usize, usize, Option<T>)> {
+        match &mut self.inner {
+            Impl::Empty(_size) => Err(EINVAL),
+            Impl::Array(array) => array.reserve_existing(offset),
+            Impl::Tree(tree) => tree.reserve_existing(offset),
+        }
+    }
+
+    /// Call the provided callback at every allocated region.
+    ///
+    /// This destroys the range allocator. Used only during shutdown.
+    pub(crate) fn take_for_each<F: Fn(usize, usize, usize, Option<T>)>(&mut self, callback: F) {
+        match &mut self.inner {
+            Impl::Empty(_size) => {}
+            Impl::Array(array) => array.take_for_each(callback),
+            Impl::Tree(tree) => tree.take_for_each(callback),
+        }
+    }
+}
+
+/// The arguments for `reserve_new`.
+#[derive(Default)]
+pub(crate) struct ReserveNewArgs<T> {
+    pub(crate) size: usize,
+    pub(crate) is_oneway: bool,
+    pub(crate) debug_id: usize,
+    pub(crate) pid: Pid,
+    pub(crate) empty_array_alloc: Option<EmptyArrayAlloc<T>>,
+    pub(crate) new_tree_alloc: Option<FromArrayAllocs<T>>,
+    pub(crate) tree_alloc: Option<ReserveNewTreeAlloc<T>>,
+}
+
+/// The return type of `ReserveNew`.
+pub(crate) enum ReserveNew<T> {
+    Success(ReserveNewSuccess<T>),
+    NeedAlloc(ReserveNewNeedAlloc<T>),
+}
+
+/// Returned by `reserve_new` when the reservation was successul.
+pub(crate) struct ReserveNewSuccess<T> {
+    pub(crate) offset: usize,
+    pub(crate) oneway_spam_detected: bool,
+
+    // If the user supplied an allocation that we did not end up using, then we return it here.
+    // The caller will kfree it outside of the lock.
+    _empty_array_alloc: Option<EmptyArrayAlloc<T>>,
+    _new_tree_alloc: Option<FromArrayAllocs<T>>,
+    _tree_alloc: Option<ReserveNewTreeAlloc<T>>,
+}
+
+/// Returned by `reserve_new` to request the caller to make an allocation before calling the method
+/// again.
+pub(crate) struct ReserveNewNeedAlloc<T> {
+    args: ReserveNewArgs<T>,
+    need_empty_array_alloc: bool,
+    need_new_tree_alloc: bool,
+    need_tree_alloc: bool,
+}
+
+impl<T> ReserveNewNeedAlloc<T> {
+    /// Make the necessary allocations for another call to `reserve_new`.
+    pub(crate) fn make_alloc(mut self) -> Result<ReserveNewArgs<T>> {
+        if self.need_empty_array_alloc && self.args.empty_array_alloc.is_none() {
+            self.args.empty_array_alloc = Some(EmptyArrayAlloc::try_new(TREE_THRESHOLD)?);
+        }
+        if self.need_new_tree_alloc && self.args.new_tree_alloc.is_none() {
+            self.args.new_tree_alloc = Some(FromArrayAllocs::try_new(TREE_THRESHOLD)?);
+        }
+        if self.need_tree_alloc && self.args.tree_alloc.is_none() {
+            self.args.tree_alloc = Some(ReserveNewTreeAlloc::try_new()?);
+        }
+        Ok(self.args)
+    }
+}

drivers/android/binder/range_alloc/tree.rs

@@ -0,0 +1,500 @@
+// SPDX-License-Identifier: GPL-2.0
+
+// Copyright (C) 2024 Google LLC.
+
+use kernel::{
+    page::PAGE_SIZE,
+    prelude::*,
+    rbtree::{RBTree, RBTreeNode, RBTreeNodeReservation},
+    seq_file::SeqFile,
+    seq_print,
+    task::Pid,
+};
+
+use crate::range_alloc::{DescriptorState, FreedRange, Range};
+
+/// Keeps track of allocations in a process' mmap.
+///
+/// Each process has an mmap where the data for incoming transactions will be placed. This struct
+/// keeps track of allocations made in the mmap. For each allocation, we store a descriptor that
+/// has metadata related to the allocation. We also keep track of available free space.
+pub(super) struct TreeRangeAllocator<T> {
+    /// This collection contains descriptors for *both* ranges containing an allocation, *and* free
+    /// ranges between allocations. The free ranges get merged, so there are never two free ranges
+    /// next to each other.
+    tree: RBTree<usize, Descriptor<T>>,
+    /// Contains an entry for every free range in `self.tree`. This tree sorts the ranges by size,
+    /// letting us look up the smallest range whose size is at least some lower bound.
+    free_tree: RBTree<FreeKey, ()>,
+    size: usize,
+    free_oneway_space: usize,
+}
+
+impl<T> TreeRangeAllocator<T> {
+    pub(crate) fn from_array(
+        size: usize,
+        ranges: &mut Vec<Range<T>>,
+        alloc: &mut FromArrayAllocs<T>,
+    ) -> Self {
+        let mut tree = TreeRangeAllocator {
+            tree: RBTree::new(),
+            free_tree: RBTree::new(),
+            size,
+            free_oneway_space: size / 2,
+        };
+
+        let mut free_offset = 0;
+        for range in ranges.drain(..) {
+            let free_size = range.offset - free_offset;
+            if free_size > 0 {
+                let free_node = alloc.free_tree.pop().unwrap();
+                tree.free_tree
+                    .insert(free_node.into_node((free_size, free_offset), ()));
+                let tree_node = alloc.tree.pop().unwrap();
+                tree.tree.insert(
+                    tree_node.into_node(free_offset, Descriptor::new(free_offset, free_size)),
+                );
+            }
+            free_offset = range.endpoint();
+
+            if range.state.is_oneway() {
+                tree.free_oneway_space = tree.free_oneway_space.saturating_sub(range.size);
+            }
+
+            let free_res = alloc.free_tree.pop().unwrap();
+            let tree_node = alloc.tree.pop().unwrap();
+            let mut desc = Descriptor::new(range.offset, range.size);
+            desc.state = Some((range.state, free_res));
+            tree.tree.insert(tree_node.into_node(range.offset, desc));
+        }
+
+        // After the last range, we may need a free range.
+        if free_offset < size {
+            let free_size = size - free_offset;
+            let free_node = alloc.free_tree.pop().unwrap();
+            tree.free_tree
+                .insert(free_node.into_node((free_size, free_offset), ()));
+            let tree_node = alloc.tree.pop().unwrap();
+            tree.tree
+                .insert(tree_node.into_node(free_offset, Descriptor::new(free_offset, free_size)));
+        }
+
+        tree
+    }
+
+    pub(crate) fn is_empty(&self) -> bool {
+        let mut tree_iter = self.tree.values();
+        // There's always at least one range, because index zero is either the start of a free or
+        // allocated range.
+        let first_value = tree_iter.next().unwrap();
+        if tree_iter.next().is_some() {
+            // There are never two free ranges next to each other, so if there is more than one
+            // descriptor, then at least one of them must hold an allocated range.
+            return false;
+        }
+        // There is only one descriptor. Return true if it is for a free range.
+        first_value.state.is_none()
+    }
+
+    pub(crate) fn total_size(&self) -> usize {
+        self.size
+    }
+
+    pub(crate) fn free_oneway_space(&self) -> usize {
+        self.free_oneway_space
+    }
+
+    pub(crate) fn count_buffers(&self) -> usize {
+        self.tree
+            .values()
+            .filter(|desc| desc.state.is_some())
+            .count()
+    }
+
+    pub(crate) fn debug_print(&self, m: &SeqFile) -> Result<()> {
+        for desc in self.tree.values() {
+            let state = match &desc.state {
+                Some(state) => &state.0,
+                None => continue,
+            };
+            seq_print!(
+                m,
+                "  buffer: {} size {} pid {}",
+                desc.offset,
+                desc.size,
+                state.pid(),
+            );
+            if state.is_oneway() {
+                seq_print!(m, " oneway");
+            }
+            match state {
+                DescriptorState::Reserved(_res) => {
+                    seq_print!(m, " reserved\n");
+                }
+                DescriptorState::Allocated(_alloc) => {
+                    seq_print!(m, " allocated\n");
+                }
+            }
+        }
+        Ok(())
+    }
+
+    fn find_best_match(&mut self, size: usize) -> Option<&mut Descriptor<T>> {
+        let free_cursor = self.free_tree.cursor_lower_bound(&(size, 0))?;
+        let ((_, offset), _) = free_cursor.current();
+        self.tree.get_mut(offset)
+    }
+
+    /// Try to reserve a new buffer, using the provided allocation if necessary.
+    pub(crate) fn reserve_new(
+        &mut self,
+        debug_id: usize,
+        size: usize,
+        is_oneway: bool,
+        pid: Pid,
+        alloc: ReserveNewTreeAlloc<T>,
+    ) -> Result<(usize, bool)> {
+        // Compute new value of free_oneway_space, which is set only on success.
+        let new_oneway_space = if is_oneway {
+            match self.free_oneway_space.checked_sub(size) {
+                Some(new_oneway_space) => new_oneway_space,
+                None => return Err(ENOSPC),
+            }
+        } else {
+            self.free_oneway_space
+        };
+
+        // Start detecting spammers once we have less than 20%
+        // of async space left (which is less than 10% of total
+        // buffer size).
+        //
+        // (This will short-circut, so `low_oneway_space` is
+        // only called when necessary.)
+        let oneway_spam_detected =
+            is_oneway && new_oneway_space < self.size / 10 && self.low_oneway_space(pid);
+
+        let (found_size, found_off, tree_node, free_tree_node) = match self.find_best_match(size) {
+            None => {
+                pr_warn!("ENOSPC from range_alloc.reserve_new - size: {}", size);
+                return Err(ENOSPC);
+            }
+            Some(desc) => {
+                let found_size = desc.size;
+                let found_offset = desc.offset;
+
+                // In case we need to break up the descriptor
+                let new_desc = Descriptor::new(found_offset + size, found_size - size);
+                let (tree_node, free_tree_node, desc_node_res) = alloc.initialize(new_desc);
+
+                desc.state = Some((
+                    DescriptorState::new(is_oneway, debug_id, pid),
+                    desc_node_res,
+                ));
+                desc.size = size;
+
+                (found_size, found_offset, tree_node, free_tree_node)
+            }
+        };
+        self.free_oneway_space = new_oneway_space;
+        self.free_tree.remove(&(found_size, found_off));
+
+        if found_size != size {
+            self.tree.insert(tree_node);
+            self.free_tree.insert(free_tree_node);
+        }
+
+        Ok((found_off, oneway_spam_detected))
+    }
+
+    pub(crate) fn reservation_abort(&mut self, offset: usize) -> Result<FreedRange> {
+        let mut cursor = self.tree.cursor_lower_bound(&offset).ok_or_else(|| {
+            pr_warn!(
+                "EINVAL from range_alloc.reservation_abort - offset: {}",
+                offset
+            );
+            EINVAL
+        })?;
+
+        let (_, desc) = cursor.current_mut();
+
+        if desc.offset != offset {
+            pr_warn!(
+                "EINVAL from range_alloc.reservation_abort - offset: {}",
+                offset
+            );
+            return Err(EINVAL);
+        }
+
+        let (reservation, free_node_res) = desc.try_change_state(|state| match state {
+            Some((DescriptorState::Reserved(reservation), free_node_res)) => {
+                (None, Ok((reservation, free_node_res)))
+            }
+            None => {
+                pr_warn!(
+                    "EINVAL from range_alloc.reservation_abort - offset: {}",
+                    offset
+                );
+                (None, Err(EINVAL))
+            }
+            allocated => {
+                pr_warn!(
+                    "EPERM from range_alloc.reservation_abort - offset: {}",
+                    offset
+                );
+                (allocated, Err(EPERM))
+            }
+        })?;
+
+        let mut size = desc.size;
+        let mut offset = desc.offset;
+        let free_oneway_space_add = if reservation.is_oneway { size } else { 0 };
+
+        self.free_oneway_space += free_oneway_space_add;
+
+        let mut freed_range = FreedRange::interior_pages(offset, size);
+        // Compute how large the next free region needs to be to include one more page in
+        // the newly freed range.
+        let add_next_page_needed = match (offset + size) % PAGE_SIZE {
+            0 => usize::MAX,
+            unalign => PAGE_SIZE - unalign,
+        };
+        // Compute how large the previous free region needs to be to include one more page
+        // in the newly freed range.
+        let add_prev_page_needed = match offset % PAGE_SIZE {
+            0 => usize::MAX,
+            unalign => unalign,
+        };
+
+        // Merge next into current if next is free
+        let remove_next = match cursor.peek_next() {
+            Some((_, next)) if next.state.is_none() => {
+                if next.size >= add_next_page_needed {
+                    freed_range.end_page_idx += 1;
+                }
+                self.free_tree.remove(&(next.size, next.offset));
+                size += next.size;
+                true
+            }
+            _ => false,
+        };
+
+        if remove_next {
+            let (_, desc) = cursor.current_mut();
+            desc.size = size;
+            cursor.remove_next();
+        }
+
+        // Merge current into prev if prev is free
+        match cursor.peek_prev_mut() {
+            Some((_, prev)) if prev.state.is_none() => {
+                if prev.size >= add_prev_page_needed {
+                    freed_range.start_page_idx -= 1;
+                }
+                // merge previous with current, remove current
+                self.free_tree.remove(&(prev.size, prev.offset));
+                offset = prev.offset;
+                size += prev.size;
+                prev.size = size;
+                cursor.remove_current();
+            }
+            _ => {}
+        };
+
+        self.free_tree
+            .insert(free_node_res.into_node((size, offset), ()));
+
+        Ok(freed_range)
+    }
+
+    pub(crate) fn reservation_commit(&mut self, offset: usize, data: Option<T>) -> Result {
+        let desc = self.tree.get_mut(&offset).ok_or_else(|| {
+            pr_warn!(
+                "ENOENT from range_alloc.reservation_commit - offset: {}",
+                offset
+            );
+            ENOENT
+        })?;
+
+        desc.try_change_state(|state| match state {
+            Some((DescriptorState::Reserved(reservation), free_node_res)) => (
+                Some((
+                    DescriptorState::Allocated(reservation.allocate(data)),
+                    free_node_res,
+                )),
+                Ok(()),
+            ),
+            other => {
+                pr_warn!(
+                    "ENOENT from range_alloc.reservation_commit - offset: {}",
+                    offset
+                );
+                (other, Err(ENOENT))
+            }
+        })
+    }
+
+    /// Takes an entry at the given offset from [`DescriptorState::Allocated`] to
+    /// [`DescriptorState::Reserved`].
+    ///
+    /// Returns the size of the existing entry and the data associated with it.
+    pub(crate) fn reserve_existing(&mut self, offset: usize) -> Result<(usize, usize, Option<T>)> {
+        let desc = self.tree.get_mut(&offset).ok_or_else(|| {
+            pr_warn!(
+                "ENOENT from range_alloc.reserve_existing - offset: {}",
+                offset
+            );
+            ENOENT
+        })?;
+
+        let (debug_id, data) = desc.try_change_state(|state| match state {
+            Some((DescriptorState::Allocated(allocation), free_node_res)) => {
+                let (reservation, data) = allocation.deallocate();
+                let debug_id = reservation.debug_id;
+                (
+                    Some((DescriptorState::Reserved(reservation), free_node_res)),
+                    Ok((debug_id, data)),
+                )
+            }
+            other => {
+                pr_warn!(
+                    "ENOENT from range_alloc.reserve_existing - offset: {}",
+                    offset
+                );
+                (other, Err(ENOENT))
+            }
+        })?;
+
+        Ok((desc.size, debug_id, data))
+    }
+
+    /// Call the provided callback at every allocated region.
+    ///
+    /// This destroys the range allocator. Used only during shutdown.
+    pub(crate) fn take_for_each<F: Fn(usize, usize, usize, Option<T>)>(&mut self, callback: F) {
+        for (_, desc) in self.tree.iter_mut() {
+            if let Some((DescriptorState::Allocated(allocation), _)) = &mut desc.state {
+                callback(
+                    desc.offset,
+                    desc.size,
+                    allocation.debug_id(),
+                    allocation.take(),
+                );
+            }
+        }
+    }
+
+    /// Find the amount and size of buffers allocated by the current caller.
+    ///
+    /// The idea is that once we cross the threshold, whoever is responsible
+    /// for the low async space is likely to try to send another async transaction,
+    /// and at some point we'll catch them in the act.  This is more efficient
+    /// than keeping a map per pid.
+    fn low_oneway_space(&self, calling_pid: Pid) -> bool {
+        let mut total_alloc_size = 0;
+        let mut num_buffers = 0;
+        for (_, desc) in self.tree.iter() {
+            if let Some((state, _)) = &desc.state {
+                if state.is_oneway() && state.pid() == calling_pid {
+                    total_alloc_size += desc.size;
+                    num_buffers += 1;
+                }
+            }
+        }
+
+        // Warn if this pid has more than 50 transactions, or more than 50% of
+        // async space (which is 25% of total buffer size). Oneway spam is only
+        // detected when the threshold is exceeded.
+        num_buffers > 50 || total_alloc_size > self.size / 4
+    }
+}
+
+type TreeDescriptorState<T> = (DescriptorState<T>, FreeNodeRes);
+struct Descriptor<T> {
+    size: usize,
+    offset: usize,
+    state: Option<TreeDescriptorState<T>>,
+}
+
+impl<T> Descriptor<T> {
+    fn new(offset: usize, size: usize) -> Self {
+        Self {
+            size,
+            offset,
+            state: None,
+        }
+    }
+
+    fn try_change_state<F, Data>(&mut self, f: F) -> Result<Data>
+    where
+        F: FnOnce(Option<TreeDescriptorState<T>>) -> (Option<TreeDescriptorState<T>>, Result<Data>),
+    {
+        let (new_state, result) = f(self.state.take());
+        self.state = new_state;
+        result
+    }
+}
+
+// (Descriptor.size, Descriptor.offset)
+type FreeKey = (usize, usize);
+type FreeNodeRes = RBTreeNodeReservation<FreeKey, ()>;
+
+/// An allocation for use by `reserve_new`.
+pub(crate) struct ReserveNewTreeAlloc<T> {
+    tree_node_res: RBTreeNodeReservation<usize, Descriptor<T>>,
+    free_tree_node_res: FreeNodeRes,
+    desc_node_res: FreeNodeRes,
+}
+
+impl<T> ReserveNewTreeAlloc<T> {
+    pub(crate) fn try_new() -> Result<Self> {
+        let tree_node_res = RBTreeNodeReservation::new(GFP_KERNEL)?;
+        let free_tree_node_res = RBTreeNodeReservation::new(GFP_KERNEL)?;
+        let desc_node_res = RBTreeNodeReservation::new(GFP_KERNEL)?;
+        Ok(Self {
+            tree_node_res,
+            free_tree_node_res,
+            desc_node_res,
+        })
+    }
+
+    fn initialize(
+        self,
+        desc: Descriptor<T>,
+    ) -> (
+        RBTreeNode<usize, Descriptor<T>>,
+        RBTreeNode<FreeKey, ()>,
+        FreeNodeRes,
+    ) {
+        let size = desc.size;
+        let offset = desc.offset;
+        (
+            self.tree_node_res.into_node(offset, desc),
+            self.free_tree_node_res.into_node((size, offset), ()),
+            self.desc_node_res,
+        )
+    }
+}
+
+/// An allocation for creating a tree from an `ArrayRangeAllocator`.
+pub(crate) struct FromArrayAllocs<T> {
+    tree: Vec<RBTreeNodeReservation<usize, Descriptor<T>>>,
+    free_tree: Vec<RBTreeNodeReservation<FreeKey, ()>>,
+}
+
+impl<T> FromArrayAllocs<T> {
+    pub(crate) fn try_new(len: usize) -> Result<Self> {
+        let num_descriptors = 2 * len + 1;
+
+        let mut tree = Vec::with_capacity(num_descriptors, GFP_KERNEL)?;
+        for _ in 0..num_descriptors {
+            tree.push(RBTreeNodeReservation::new(GFP_KERNEL)?, GFP_KERNEL)?;
+        }
+
+        let mut free_tree = Vec::with_capacity(num_descriptors, GFP_KERNEL)?;
+        for _ in 0..num_descriptors {
+            free_tree.push(RBTreeNodeReservation::new(GFP_KERNEL)?, GFP_KERNEL)?;
+        }
+
+        Ok(Self { tree, free_tree })
+    }
+}

drivers/android/binder/rust_binder.rs

@@ -0,0 +1,618 @@
+// SPDX-License-Identifier: GPL-2.0
+
+// Copyright (C) 2024 Google LLC.
+
+//! Binder -- the Android IPC mechanism.
+#![recursion_limit = "256"]
+
+use kernel::{
+    bindings::{self, seq_file},
+    fs::File,
+    list::{HasListLinks, ListArc, ListArcSafe, ListLinksSelfPtr, TryNewListArc},
+    prelude::*,
+    seq_file::SeqFile,
+    seq_print,
+    sync::poll::PollTable,
+    sync::Arc,
+    task::Pid,
+    types::{AsBytes, ForeignOwnable},
+    uaccess::UserSliceWriter,
+};
+
+use crate::{context::Context, page_range::Shrinker, process::Process, thread::Thread};
+
+use core::{
+    ptr::NonNull,
+    sync::atomic::{AtomicBool, AtomicUsize, Ordering},
+};
+
+mod allocation;
+mod context;
+mod deferred_close;
+mod defs;
+mod error;
+mod node;
+mod page_range;
+mod prio;
+mod process;
+mod range_alloc;
+mod stats;
+mod thread;
+mod trace;
+mod transaction;
+
+#[allow(warnings)] // generated bindgen code
+mod binderfs {
+    use kernel::bindings::{dentry, inode};
+
+    extern "C" {
+        pub fn init_rust_binderfs() -> core::ffi::c_int;
+    }
+    extern "C" {
+        pub fn rust_binderfs_create_proc_file(
+            nodp: *mut inode,
+            pid: core::ffi::c_int,
+        ) -> *mut dentry;
+    }
+    extern "C" {
+        pub fn rust_binderfs_remove_file(dentry: *mut dentry);
+    }
+    pub type rust_binder_context = *mut core::ffi::c_void;
+    #[repr(C)]
+    #[derive(Copy, Clone)]
+    pub struct binder_device {
+        pub minor: core::ffi::c_int,
+        pub ctx: rust_binder_context,
+    }
+    impl Default for binder_device {
+        fn default() -> Self {
+            let mut s = ::core::mem::MaybeUninit::<Self>::uninit();
+            unsafe {
+                ::core::ptr::write_bytes(s.as_mut_ptr(), 0, 1);
+                s.assume_init()
+            }
+        }
+    }
+}
+
+module! {
+    type: BinderModule,
+    name: "rust_binder",
+    author: "Wedson Almeida Filho, Alice Ryhl",
+    description: "Android Binder",
+    license: "GPL",
+}
+
+fn next_debug_id() -> usize {
+    static NEXT_DEBUG_ID: AtomicUsize = AtomicUsize::new(0);
+
+    NEXT_DEBUG_ID.fetch_add(1, Ordering::Relaxed)
+}
+
+/// Provides a single place to write Binder return values via the
+/// supplied `UserSliceWriter`.
+pub(crate) struct BinderReturnWriter<'a> {
+    writer: UserSliceWriter,
+    thread: &'a Thread,
+}
+
+impl<'a> BinderReturnWriter<'a> {
+    fn new(writer: UserSliceWriter, thread: &'a Thread) -> Self {
+        BinderReturnWriter { writer, thread }
+    }
+
+    /// Write a return code back to user space.
+    /// Should be a `BR_` constant from [`defs`] e.g. [`defs::BR_TRANSACTION_COMPLETE`].
+    fn write_code(&mut self, code: u32) -> Result {
+        crate::trace::trace_return(code);
+        stats::GLOBAL_STATS.inc_br(code);
+        self.thread.process.stats.inc_br(code);
+        self.writer.write(&code)
+    }
+
+    /// Write something *other than* a return code to user space.
+    fn write_payload<T: AsBytes>(&mut self, payload: &T) -> Result {
+        self.writer.write(payload)
+    }
+
+    fn len(&self) -> usize {
+        self.writer.len()
+    }
+}
+
+/// Specifies how a type should be delivered to the read part of a BINDER_WRITE_READ ioctl.
+///
+/// When a value is pushed to the todo list for a process or thread, it is stored as a trait object
+/// with the type `Arc<dyn DeliverToRead>`. Trait objects are a Rust feature that lets you
+/// implement dynamic dispatch over many different types. This lets us store many different types
+/// in the todo list.
+trait DeliverToRead: ListArcSafe + Send + Sync {
+    /// Performs work. Returns true if remaining work items in the queue should be processed
+    /// immediately, or false if it should return to caller before processing additional work
+    /// items.
+    fn do_work(
+        self: DArc<Self>,
+        thread: &Thread,
+        writer: &mut BinderReturnWriter<'_>,
+    ) -> Result<bool>;
+
+    /// Cancels the given work item. This is called instead of [`DeliverToRead::do_work`] when work
+    /// won't be delivered.
+    fn cancel(self: DArc<Self>);
+
+    /// Called when a work item is delivered directly to a specific thread, rather than to the
+    /// process work list.
+    fn on_thread_selected(&self, _thread: &thread::Thread);
+
+    /// Should we use `wake_up_interruptible_sync` or `wake_up_interruptible` when scheduling this
+    /// work item?
+    ///
+    /// Generally only set to true for non-oneway transactions.
+    fn should_sync_wakeup(&self) -> bool;
+
+    fn debug_print(&self, m: &SeqFile, prefix: &str, transaction_prefix: &str) -> Result<()>;
+}
+
+// Wrapper around a `DeliverToRead` with linked list links.
+#[pin_data]
+struct DTRWrap<T: ?Sized> {
+    #[pin]
+    links: ListLinksSelfPtr<DTRWrap<dyn DeliverToRead>>,
+    #[pin]
+    wrapped: T,
+}
+kernel::list::impl_has_list_links_self_ptr! {
+    impl HasSelfPtr<DTRWrap<dyn DeliverToRead>> for DTRWrap<dyn DeliverToRead> { self.links }
+}
+kernel::list::impl_list_arc_safe! {
+    impl{T: ListArcSafe + ?Sized} ListArcSafe<0> for DTRWrap<T> {
+        tracked_by wrapped: T;
+    }
+}
+kernel::list::impl_list_item! {
+    impl ListItem<0> for DTRWrap<dyn DeliverToRead> {
+        using ListLinksSelfPtr;
+    }
+}
+
+impl<T: ?Sized> core::ops::Deref for DTRWrap<T> {
+    type Target = T;
+    fn deref(&self) -> &T {
+        &self.wrapped
+    }
+}
+
+type DArc<T> = kernel::sync::Arc<DTRWrap<T>>;
+type DLArc<T> = kernel::list::ListArc<DTRWrap<T>>;
+
+impl<T: ListArcSafe> DTRWrap<T> {
+    fn new(val: impl PinInit<T>) -> impl PinInit<Self> {
+        pin_init!(Self {
+            links <- ListLinksSelfPtr::new(),
+            wrapped <- val,
+        })
+    }
+
+    fn arc_try_new(val: T) -> Result<DLArc<T>, kernel::alloc::AllocError> {
+        ListArc::pin_init(
+            try_pin_init!(Self {
+                links <- ListLinksSelfPtr::new(),
+                wrapped: val,
+            }),
+            GFP_KERNEL,
+        )
+        .map_err(|_| kernel::alloc::AllocError)
+    }
+
+    fn arc_pin_init(init: impl PinInit<T>) -> Result<DLArc<T>, kernel::error::Error> {
+        ListArc::pin_init(
+            try_pin_init!(Self {
+                links <- ListLinksSelfPtr::new(),
+                wrapped <- init,
+            }),
+            GFP_KERNEL,
+        )
+    }
+}
+
+struct DeliverCode {
+    code: u32,
+    skip: AtomicBool,
+}
+
+kernel::list::impl_list_arc_safe! {
+    impl ListArcSafe<0> for DeliverCode { untracked; }
+}
+
+impl DeliverCode {
+    fn new(code: u32) -> Self {
+        Self {
+            code,
+            skip: AtomicBool::new(false),
+        }
+    }
+
+    /// Disable this DeliverCode and make it do nothing.
+    ///
+    /// This is used instead of removing it from the work list, since `LinkedList::remove` is
+    /// unsafe, whereas this method is not.
+    fn skip(&self) {
+        self.skip.store(true, Ordering::Relaxed);
+    }
+}
+
+impl DeliverToRead for DeliverCode {
+    fn do_work(
+        self: DArc<Self>,
+        _thread: &Thread,
+        writer: &mut BinderReturnWriter<'_>,
+    ) -> Result<bool> {
+        if !self.skip.load(Ordering::Relaxed) {
+            writer.write_code(self.code)?;
+        }
+        Ok(true)
+    }
+
+    fn cancel(self: DArc<Self>) {}
+    fn on_thread_selected(&self, _thread: &Thread) {}
+
+    fn should_sync_wakeup(&self) -> bool {
+        false
+    }
+
+    fn debug_print(&self, m: &SeqFile, prefix: &str, _tprefix: &str) -> Result<()> {
+        seq_print!(m, "{}", prefix);
+        if self.skip.load(Ordering::Relaxed) {
+            seq_print!(m, "(skipped) ");
+        }
+        if self.code == defs::BR_TRANSACTION_COMPLETE {
+            seq_print!(m, "transaction complete\n");
+        } else {
+            seq_print!(m, "transaction error: {}\n", self.code);
+        }
+        Ok(())
+    }
+}
+
+const fn ptr_align(value: usize) -> usize {
+    let size = core::mem::size_of::<usize>() - 1;
+    (value + size) & !size
+}
+
+// SAFETY: We call register in `init`.
+static BINDER_SHRINKER: Shrinker = unsafe { Shrinker::new() };
+
+struct BinderModule {}
+
+impl kernel::Module for BinderModule {
+    fn init(_module: &'static kernel::ThisModule) -> Result<Self> {
+        // SAFETY: The module initializer never runs twice, so we only call this once.
+        unsafe { crate::context::CONTEXTS.init() };
+
+        // SAFETY: This just accesses global booleans.
+        unsafe {
+            extern "C" {
+                static mut binder_use_rust: i32;
+                fn unload_binder() -> i32;
+            }
+
+            if binder_use_rust == 0 {
+                return Ok(Self {});
+            }
+            if unload_binder() != 0 {
+                pr_err!("Failed to unload C Binder.");
+                return Ok(Self {});
+            }
+        }
+
+        BINDER_SHRINKER.register(kernel::c_str!("android-binder"))?;
+
+        // SAFETY: The module is being loaded, so we can initialize binderfs.
+        unsafe { kernel::error::to_result(binderfs::init_rust_binderfs())? };
+
+        Ok(Self {})
+    }
+}
+
+/// Makes the inner type Sync.
+#[repr(transparent)]
+pub struct AssertSync<T>(T);
+// SAFETY: Used only to insert `file_operations` into a global, which is safe.
+unsafe impl<T> Sync for AssertSync<T> {}
+
+/// File operations that rust_binderfs.c can use.
+#[no_mangle]
+#[used]
+pub static rust_binder_fops: AssertSync<kernel::bindings::file_operations> = {
+    // SAFETY: All zeroes is safe for the `file_operations` type.
+    let zeroed_ops = unsafe { core::mem::MaybeUninit::zeroed().assume_init() };
+
+    let ops = kernel::bindings::file_operations {
+        owner: THIS_MODULE.as_ptr(),
+        poll: Some(rust_binder_poll),
+        unlocked_ioctl: Some(rust_binder_unlocked_ioctl),
+        compat_ioctl: Some(rust_binder_compat_ioctl),
+        mmap: Some(rust_binder_mmap),
+        open: Some(rust_binder_open),
+        release: Some(rust_binder_release),
+        flush: Some(rust_binder_flush),
+        ..zeroed_ops
+    };
+    AssertSync(ops)
+};
+
+#[no_mangle]
+unsafe extern "C" fn rust_binder_new_context(
+    name: *const core::ffi::c_char,
+) -> *mut core::ffi::c_void {
+    // SAFETY: The caller will always provide a valid c string here.
+    let name = unsafe { kernel::str::CStr::from_char_ptr(name) };
+    match Context::new(name) {
+        Ok(ctx) => Arc::into_foreign(ctx).cast_mut(),
+        Err(_err) => core::ptr::null_mut(),
+    }
+}
+
+#[no_mangle]
+unsafe extern "C" fn rust_binder_remove_context(device: *mut core::ffi::c_void) {
+    if !device.is_null() {
+        // SAFETY: The caller ensures that the `device` pointer came from a previous call to
+        // `rust_binder_new_device`.
+        let ctx = unsafe { Arc::<Context>::from_foreign(device) };
+        ctx.deregister();
+        drop(ctx);
+    }
+}
+
+unsafe extern "C" fn rust_binder_open(
+    inode: *mut bindings::inode,
+    file_ptr: *mut bindings::file,
+) -> core::ffi::c_int {
+    // SAFETY: The `rust_binderfs.c` file ensures that `i_private` is set to a
+    // `struct binder_device`.
+    let device = unsafe { (*inode).i_private } as *const binderfs::binder_device;
+
+    assert!(!device.is_null());
+
+    // SAFETY: The `rust_binderfs.c` file ensures that `device->ctx` holds a binder context when
+    // using the rust binder fops.
+    let ctx = unsafe { Arc::<Context>::borrow((*device).ctx) };
+
+    // SAFETY: The caller provides a valid file pointer to a new `struct file`.
+    let file = unsafe { File::from_raw_file(file_ptr) };
+    let process = match Process::open(ctx, file) {
+        Ok(process) => process,
+        Err(err) => return err.to_errno(),
+    };
+
+    // SAFETY: This is an `inode` for a newly created binder file.
+    match unsafe { BinderfsProcFile::new(inode, process.task.pid()) } {
+        Ok(Some(file)) => process.inner.lock().binderfs_file = Some(file),
+        Ok(None) => { /* pid already exists */ }
+        Err(err) => return err.to_errno(),
+    }
+
+    // SAFETY: This file is associated with Rust binder, so we own the `private_data` field.
+    unsafe { (*file_ptr).private_data = process.into_foreign().cast_mut() };
+    0
+}
+
+unsafe extern "C" fn rust_binder_release(
+    _inode: *mut bindings::inode,
+    file: *mut bindings::file,
+) -> core::ffi::c_int {
+    // SAFETY: We previously set `private_data` in `rust_binder_open`.
+    let process = unsafe { Arc::<Process>::from_foreign((*file).private_data) };
+    // SAFETY: The caller ensures that the file is valid.
+    let file = unsafe { File::from_raw_file(file) };
+    Process::release(process, file);
+    0
+}
+
+unsafe extern "C" fn rust_binder_compat_ioctl(
+    file: *mut bindings::file,
+    cmd: core::ffi::c_uint,
+    arg: core::ffi::c_ulong,
+) -> core::ffi::c_long {
+    // SAFETY: We previously set `private_data` in `rust_binder_open`.
+    let f = unsafe { Arc::<Process>::borrow((*file).private_data) };
+    // SAFETY: The caller ensures that the file is valid.
+    match Process::compat_ioctl(f, unsafe { File::from_raw_file(file) }, cmd as _, arg as _) {
+        Ok(()) => 0,
+        Err(err) => err.to_errno().into(),
+    }
+}
+
+unsafe extern "C" fn rust_binder_unlocked_ioctl(
+    file: *mut bindings::file,
+    cmd: core::ffi::c_uint,
+    arg: core::ffi::c_ulong,
+) -> core::ffi::c_long {
+    // SAFETY: We previously set `private_data` in `rust_binder_open`.
+    let f = unsafe { Arc::<Process>::borrow((*file).private_data) };
+    // SAFETY: The caller ensures that the file is valid.
+    match Process::ioctl(f, unsafe { File::from_raw_file(file) }, cmd as _, arg as _) {
+        Ok(()) => 0,
+        Err(err) => err.to_errno().into(),
+    }
+}
+
+unsafe extern "C" fn rust_binder_mmap(
+    file: *mut bindings::file,
+    vma: *mut bindings::vm_area_struct,
+) -> core::ffi::c_int {
+    // SAFETY: We previously set `private_data` in `rust_binder_open`.
+    let f = unsafe { Arc::<Process>::borrow((*file).private_data) };
+    // SAFETY: The caller ensures that the vma is valid.
+    let area = unsafe { kernel::mm::virt::VmAreaNew::from_raw(vma) };
+    // SAFETY: The caller ensures that the file is valid.
+    match Process::mmap(f, unsafe { File::from_raw_file(file) }, area) {
+        Ok(()) => 0,
+        Err(err) => err.to_errno(),
+    }
+}
+
+unsafe extern "C" fn rust_binder_poll(
+    file: *mut bindings::file,
+    wait: *mut bindings::poll_table_struct,
+) -> bindings::__poll_t {
+    // SAFETY: We previously set `private_data` in `rust_binder_open`.
+    let f = unsafe { Arc::<Process>::borrow((*file).private_data) };
+    // SAFETY: The caller ensures that the file is valid.
+    let fileref = unsafe { File::from_raw_file(file) };
+    // SAFETY: The caller ensures that the `PollTable` is valid.
+    match Process::poll(f, fileref, unsafe { PollTable::from_ptr(wait) }) {
+        Ok(v) => v,
+        Err(_) => bindings::POLLERR,
+    }
+}
+
+unsafe extern "C" fn rust_binder_flush(
+    file: *mut bindings::file,
+    _id: bindings::fl_owner_t,
+) -> core::ffi::c_int {
+    // SAFETY: We previously set `private_data` in `rust_binder_open`.
+    let f = unsafe { Arc::<Process>::borrow((*file).private_data) };
+    match Process::flush(f) {
+        Ok(()) => 0,
+        Err(err) => err.to_errno(),
+    }
+}
+
+#[no_mangle]
+unsafe extern "C" fn rust_binder_stats_show(
+    ptr: *mut seq_file,
+    _: *mut core::ffi::c_void,
+) -> core::ffi::c_int {
+    // SAFETY: The caller ensures that the pointer is valid and exclusive for the duration in which
+    // this method is called.
+    let m = unsafe { SeqFile::from_raw(ptr) };
+    if let Err(err) = rust_binder_stats_show_impl(m) {
+        seq_print!(m, "failed to generate state: {:?}\n", err);
+    }
+    0
+}
+
+#[no_mangle]
+unsafe extern "C" fn rust_binder_state_show(
+    ptr: *mut seq_file,
+    _: *mut core::ffi::c_void,
+) -> core::ffi::c_int {
+    // SAFETY: The caller ensures that the pointer is valid and exclusive for the duration in which
+    // this method is called.
+    let m = unsafe { SeqFile::from_raw(ptr) };
+    if let Err(err) = rust_binder_state_show_impl(m) {
+        seq_print!(m, "failed to generate state: {:?}\n", err);
+    }
+    0
+}
+
+#[no_mangle]
+unsafe extern "C" fn rust_binder_proc_show(
+    ptr: *mut seq_file,
+    _: *mut core::ffi::c_void,
+) -> core::ffi::c_int {
+    // SAFETY: Accessing the private field of `seq_file` is okay.
+    let pid = (unsafe { (*ptr).private }) as usize as Pid;
+    // SAFETY: The caller ensures that the pointer is valid and exclusive for the duration in which
+    // this method is called.
+    let m = unsafe { SeqFile::from_raw(ptr) };
+    if let Err(err) = rust_binder_proc_show_impl(m, pid) {
+        seq_print!(m, "failed to generate state: {:?}\n", err);
+    }
+    0
+}
+
+#[no_mangle]
+unsafe extern "C" fn rust_binder_transactions_show(
+    ptr: *mut seq_file,
+    _: *mut core::ffi::c_void,
+) -> core::ffi::c_int {
+    // SAFETY: The caller ensures that the pointer is valid and exclusive for the duration in which
+    // this method is called.
+    let m = unsafe { SeqFile::from_raw(ptr) };
+    if let Err(err) = rust_binder_transactions_show_impl(m) {
+        seq_print!(m, "failed to generate state: {:?}\n", err);
+    }
+    0
+}
+
+fn rust_binder_transactions_show_impl(m: &SeqFile) -> Result<()> {
+    seq_print!(m, "binder transactions:\n");
+    let contexts = context::get_all_contexts()?;
+    for ctx in contexts {
+        let procs = ctx.get_all_procs()?;
+        for proc in procs {
+            proc.debug_print(m, &ctx, false)?;
+            seq_print!(m, "\n");
+        }
+    }
+    Ok(())
+}
+
+fn rust_binder_stats_show_impl(m: &SeqFile) -> Result<()> {
+    seq_print!(m, "binder stats:\n");
+    stats::GLOBAL_STATS.debug_print("", m);
+    let contexts = context::get_all_contexts()?;
+    for ctx in contexts {
+        let procs = ctx.get_all_procs()?;
+        for proc in procs {
+            proc.debug_print_stats(m, &ctx)?;
+            seq_print!(m, "\n");
+        }
+    }
+    Ok(())
+}
+
+fn rust_binder_state_show_impl(m: &SeqFile) -> Result<()> {
+    seq_print!(m, "binder state:\n");
+    let contexts = context::get_all_contexts()?;
+    for ctx in contexts {
+        let procs = ctx.get_all_procs()?;
+        for proc in procs {
+            proc.debug_print(m, &ctx, true)?;
+            seq_print!(m, "\n");
+        }
+    }
+    Ok(())
+}
+
+fn rust_binder_proc_show_impl(m: &SeqFile, pid: Pid) -> Result<()> {
+    seq_print!(m, "binder proc state:\n");
+    let contexts = context::get_all_contexts()?;
+    for ctx in contexts {
+        let procs = ctx.get_procs_with_pid(pid)?;
+        for proc in procs {
+            proc.debug_print(m, &ctx, true)?;
+            seq_print!(m, "\n");
+        }
+    }
+    Ok(())
+}
+
+struct BinderfsProcFile(NonNull<bindings::dentry>);
+
+// SAFETY: Safe to drop any thread.
+unsafe impl Send for BinderfsProcFile {}
+
+impl BinderfsProcFile {
+    /// # Safety
+    ///
+    /// Takes an inode from a newly created binder file.
+    unsafe fn new(nodp: *mut bindings::inode, pid: i32) -> Result<Option<Self>> {
+        // SAFETY: The caller passes an `inode` for a newly created binder file.
+        let dentry = unsafe { binderfs::rust_binderfs_create_proc_file(nodp, pid) };
+        match kernel::error::from_err_ptr(dentry) {
+            Ok(dentry) => Ok(NonNull::new(dentry).map(Self)),
+            Err(err) if err == EEXIST => Ok(None),
+            Err(err) => Err(err),
+        }
+    }
+}
+
+impl Drop for BinderfsProcFile {
+    fn drop(&mut self) {
+        // SAFETY: This is a dentry from `rust_binderfs_remove_file` that has not been deleted yet.
+        unsafe { binderfs::rust_binderfs_remove_file(self.0.as_ptr()) };
+    }
+}

drivers/android/binder/rust_binder_events.c

@@ -0,0 +1,53 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/* rust_binder_events.c
+ *
+ * Rust Binder tracepoints.
+ *
+ * Copyright 2024 Google LLC
+ */
+
+const char * const binder_command_strings[] = {
+	"BC_TRANSACTION",
+	"BC_REPLY",
+	"BC_ACQUIRE_RESULT",
+	"BC_FREE_BUFFER",
+	"BC_INCREFS",
+	"BC_ACQUIRE",
+	"BC_RELEASE",
+	"BC_DECREFS",
+	"BC_INCREFS_DONE",
+	"BC_ACQUIRE_DONE",
+	"BC_ATTEMPT_ACQUIRE",
+	"BC_REGISTER_LOOPER",
+	"BC_ENTER_LOOPER",
+	"BC_EXIT_LOOPER",
+	"BC_REQUEST_DEATH_NOTIFICATION",
+	"BC_CLEAR_DEATH_NOTIFICATION",
+	"BC_DEAD_BINDER_DONE",
+	"BC_TRANSACTION_SG",
+	"BC_REPLY_SG",
+};
+
+const char * const binder_return_strings[] = {
+	"BR_ERROR",
+	"BR_OK",
+	"BR_TRANSACTION",
+	"BR_REPLY",
+	"BR_ACQUIRE_RESULT",
+	"BR_DEAD_REPLY",
+	"BR_TRANSACTION_COMPLETE",
+	"BR_INCREFS",
+	"BR_ACQUIRE",
+	"BR_RELEASE",
+	"BR_DECREFS",
+	"BR_ATTEMPT_ACQUIRE",
+	"BR_NOOP",
+	"BR_SPAWN_LOOPER",
+	"BR_FINISHED",
+	"BR_DEAD_BINDER",
+	"BR_CLEAR_DEATH_NOTIFICATION_DONE",
+	"BR_FAILED_REPLY",
+	"BR_FROZEN_REPLY",
+	"BR_ONEWAY_SPAM_SUSPECT",
+	"BR_TRANSACTION_PENDING_FROZEN"
+};

drivers/android/binder/rust_binder_hooks.c

@@ -0,0 +1,23 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/* rust_binder_events.c
+ *
+ * Rust Binder vendorhooks.
+ *
+ * Copyright 2024 Google LLC
+ */
+
+#include <linux/rust_binder.h>
+
+#define CREATE_TRACE_POINTS
+#define CREATE_RUST_TRACE_POINTS
+#include <trace/hooks/vendor_hooks.h>
+#include <linux/tracepoint.h>
+
+#include <trace/hooks/rust_binder.h>
+
+/*
+ * Export tracepoints that act as a bare tracehook (ie: have no trace event
+ * associated with them) to allow external modules to probe them.
+ */
+EXPORT_TRACEPOINT_SYMBOL_GPL(android_vh_rust_binder_set_priority);
+EXPORT_TRACEPOINT_SYMBOL_GPL(android_vh_rust_binder_restore_priority);

drivers/android/binder/rust_binder_internal.h

@@ -0,0 +1,87 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/* rust_binder_internal.h
+ *
+ * This file contains internal data structures used by Rust Binder. Mostly,
+ * these are type definitions used only by binderfs or things that Rust Binder
+ * define and export to binderfs.
+ *
+ * It does not include things exported by binderfs to Rust Binder since this
+ * file is not included as input to bindgen.
+ *
+ * Copyright (C) 2024 Google LLC.
+ */
+
+#ifndef _LINUX_RUST_BINDER_INTERNAL_H
+#define _LINUX_RUST_BINDER_INTERNAL_H
+
+#define RUST_BINDERFS_SUPER_MAGIC	0x6c6f6f71
+
+#include <linux/seq_file.h>
+#include <uapi/linux/android/binder.h>
+#include <uapi/linux/android/binderfs.h>
+
+/*
+ * The internal data types in the Rust Binder driver are opaque to C, so we use
+ * void pointer typedefs for these types.
+ */
+typedef void *rust_binder_context;
+
+/**
+ * struct binder_device - information about a binder device node
+ * @minor:     the minor number used by this device
+ * @ctx:       the Rust Context used by this device, or null for binder-control
+ *
+ * This is used as the private data for files directly in binderfs, but not
+ * files in the binder_logs subdirectory. This struct owns a refcount on `ctx`
+ * and the entry for `minor` in `binderfs_minors`. For binder-control `ctx` is
+ * null.
+ */
+struct binder_device {
+	int minor;
+	rust_binder_context ctx;
+};
+
+int rust_binder_stats_show(struct seq_file *m, void *unused);
+int rust_binder_state_show(struct seq_file *m, void *unused);
+int rust_binder_transactions_show(struct seq_file *m, void *unused);
+int rust_binder_proc_show(struct seq_file *m, void *pid);
+
+extern const struct file_operations rust_binder_fops;
+rust_binder_context rust_binder_new_context(char *name);
+void rust_binder_remove_context(rust_binder_context device);
+
+/**
+ * binderfs_mount_opts - mount options for binderfs
+ * @max: maximum number of allocatable binderfs binder devices
+ * @stats_mode: enable binder stats in binderfs.
+ */
+struct binderfs_mount_opts {
+	int max;
+	int stats_mode;
+};
+
+/**
+ * binderfs_info - information about a binderfs mount
+ * @ipc_ns:         The ipc namespace the binderfs mount belongs to.
+ * @control_dentry: This records the dentry of this binderfs mount
+ *                  binder-control device.
+ * @root_uid:       uid that needs to be used when a new binder device is
+ *                  created.
+ * @root_gid:       gid that needs to be used when a new binder device is
+ *                  created.
+ * @mount_opts:     The mount options in use.
+ * @device_count:   The current number of allocated binder devices.
+ * @proc_log_dir:   Pointer to the directory dentry containing process-specific
+ *                  logs.
+ */
+struct binderfs_info {
+	struct ipc_namespace *ipc_ns;
+	struct dentry *control_dentry;
+	kuid_t root_uid;
+	kgid_t root_gid;
+	struct binderfs_mount_opts mount_opts;
+	int device_count;
+	struct dentry *proc_log_dir;
+};
+
+#endif /* _LINUX_RUST_BINDER_INTERNAL_H */

drivers/android/binder/rust_binderfs.c

@@ -0,0 +1,849 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include <linux/compiler_types.h>
+#include <linux/errno.h>
+#include <linux/fs.h>
+#include <linux/fsnotify.h>
+#include <linux/gfp.h>
+#include <linux/idr.h>
+#include <linux/init.h>
+#include <linux/ipc_namespace.h>
+#include <linux/kdev_t.h>
+#include <linux/kernel.h>
+#include <linux/list.h>
+#include <linux/namei.h>
+#include <linux/magic.h>
+#include <linux/major.h>
+#include <linux/miscdevice.h>
+#include <linux/module.h>
+#include <linux/mutex.h>
+#include <linux/mount.h>
+#include <linux/fs_parser.h>
+#include <linux/sched.h>
+#include <linux/seq_file.h>
+#include <linux/slab.h>
+#include <linux/spinlock_types.h>
+#include <linux/stddef.h>
+#include <linux/string.h>
+#include <linux/types.h>
+#include <linux/uaccess.h>
+#include <linux/user_namespace.h>
+#include <linux/xarray.h>
+#include <uapi/asm-generic/errno-base.h>
+#include <uapi/linux/android/binder.h>
+#include <uapi/linux/android/binderfs.h>
+
+#include "rust_binder_internal.h"
+
+#define FIRST_INODE 1
+#define SECOND_INODE 2
+#define INODE_OFFSET 3
+#define BINDERFS_MAX_MINOR (1U << MINORBITS)
+/* Ensure that the initial ipc namespace always has devices available. */
+#define BINDERFS_MAX_MINOR_CAPPED (BINDERFS_MAX_MINOR - 4)
+
+DEFINE_SHOW_ATTRIBUTE(rust_binder_stats);
+DEFINE_SHOW_ATTRIBUTE(rust_binder_state);
+DEFINE_SHOW_ATTRIBUTE(rust_binder_transactions);
+DEFINE_SHOW_ATTRIBUTE(rust_binder_proc);
+
+char *rust_binder_devices_param = CONFIG_ANDROID_BINDER_DEVICES;
+module_param_named(rust_devices, rust_binder_devices_param, charp, 0444);
+
+static dev_t binderfs_dev;
+static DEFINE_MUTEX(binderfs_minors_mutex);
+static DEFINE_IDA(binderfs_minors);
+
+enum binderfs_param {
+	Opt_max,
+	Opt_stats_mode,
+};
+
+enum binderfs_stats_mode {
+	binderfs_stats_mode_unset,
+	binderfs_stats_mode_global,
+};
+
+struct binder_features {
+	bool oneway_spam_detection;
+	bool extended_error;
+};
+
+static const struct constant_table binderfs_param_stats[] = {
+	{ "global", binderfs_stats_mode_global },
+	{}
+};
+
+static const struct fs_parameter_spec binderfs_fs_parameters[] = {
+	fsparam_u32("max",	Opt_max),
+	fsparam_enum("stats",	Opt_stats_mode, binderfs_param_stats),
+	{}
+};
+
+static struct binder_features binder_features = {
+	.oneway_spam_detection = true,
+	.extended_error = true,
+};
+
+static inline struct binderfs_info *BINDERFS_SB(const struct super_block *sb)
+{
+	return sb->s_fs_info;
+}
+
+bool is_rust_binderfs_device(const struct inode *inode)
+{
+	if (inode->i_sb->s_magic == RUST_BINDERFS_SUPER_MAGIC)
+		return true;
+
+	return false;
+}
+
+/**
+ * binderfs_binder_device_create - allocate inode from super block of a
+ *                                 binderfs mount
+ * @ref_inode: inode from wich the super block will be taken
+ * @userp:     buffer to copy information about new device for userspace to
+ * @req:       struct binderfs_device as copied from userspace
+ *
+ * This function allocates a new binder_device and reserves a new minor
+ * number for it.
+ * Minor numbers are limited and tracked globally in binderfs_minors. The
+ * function will stash a struct binder_device for the specific binder
+ * device in i_private of the inode.
+ * It will go on to allocate a new inode from the super block of the
+ * filesystem mount, stash a struct binder_device in its i_private field
+ * and attach a dentry to that inode.
+ *
+ * Return: 0 on success, negative errno on failure
+ */
+static int binderfs_binder_device_create(struct inode *ref_inode,
+					 struct binderfs_device __user *userp,
+					 struct binderfs_device *req)
+{
+	int minor, ret;
+	struct dentry *dentry, *root;
+	struct binder_device *device = NULL;
+	rust_binder_context ctx = NULL;
+	struct inode *inode = NULL;
+	struct super_block *sb = ref_inode->i_sb;
+	struct binderfs_info *info = sb->s_fs_info;
+#if defined(CONFIG_IPC_NS)
+	bool use_reserve = (info->ipc_ns == &init_ipc_ns);
+#else
+	bool use_reserve = true;
+#endif
+
+	/* Reserve new minor number for the new device. */
+	mutex_lock(&binderfs_minors_mutex);
+	if (++info->device_count <= info->mount_opts.max)
+		minor = ida_alloc_max(&binderfs_minors,
+				      use_reserve ? BINDERFS_MAX_MINOR :
+						    BINDERFS_MAX_MINOR_CAPPED,
+				      GFP_KERNEL);
+	else
+		minor = -ENOSPC;
+	if (minor < 0) {
+		--info->device_count;
+		mutex_unlock(&binderfs_minors_mutex);
+		return minor;
+	}
+	mutex_unlock(&binderfs_minors_mutex);
+
+	ret = -ENOMEM;
+	device = kzalloc(sizeof(*device), GFP_KERNEL);
+	if (!device)
+		goto err;
+
+	req->name[BINDERFS_MAX_NAME] = '\0'; /* NUL-terminate */
+
+	ctx = rust_binder_new_context(req->name);
+	if (!ctx)
+		goto err;
+
+	inode = new_inode(sb);
+	if (!inode)
+		goto err;
+
+	inode->i_ino = minor + INODE_OFFSET;
+	simple_inode_init_ts(inode);
+	init_special_inode(inode, S_IFCHR | 0600,
+			   MKDEV(MAJOR(binderfs_dev), minor));
+	inode->i_fop = &rust_binder_fops;
+	inode->i_uid = info->root_uid;
+	inode->i_gid = info->root_gid;
+
+	req->major = MAJOR(binderfs_dev);
+	req->minor = minor;
+	device->ctx = ctx;
+	device->minor = minor;
+
+	if (userp && copy_to_user(userp, req, sizeof(*req))) {
+		ret = -EFAULT;
+		goto err;
+	}
+
+	root = sb->s_root;
+	inode_lock(d_inode(root));
+
+	/* look it up */
+	dentry = lookup_one_len(req->name, root, strlen(req->name));
+	if (IS_ERR(dentry)) {
+		inode_unlock(d_inode(root));
+		ret = PTR_ERR(dentry);
+		goto err;
+	}
+
+	if (d_really_is_positive(dentry)) {
+		/* already exists */
+		dput(dentry);
+		inode_unlock(d_inode(root));
+		ret = -EEXIST;
+		goto err;
+	}
+
+	inode->i_private = device;
+	d_instantiate(dentry, inode);
+	fsnotify_create(root->d_inode, dentry);
+	inode_unlock(d_inode(root));
+
+	return 0;
+
+err:
+	kfree(device);
+	rust_binder_remove_context(ctx);
+	mutex_lock(&binderfs_minors_mutex);
+	--info->device_count;
+	ida_free(&binderfs_minors, minor);
+	mutex_unlock(&binderfs_minors_mutex);
+	iput(inode);
+
+	return ret;
+}
+
+/**
+ * binder_ctl_ioctl - handle binder device node allocation requests
+ *
+ * The request handler for the binder-control device. All requests operate on
+ * the binderfs mount the binder-control device resides in:
+ * - BINDER_CTL_ADD
+ *   Allocate a new binder device.
+ *
+ * Return: %0 on success, negative errno on failure.
+ */
+static long binder_ctl_ioctl(struct file *file, unsigned int cmd,
+			     unsigned long arg)
+{
+	int ret = -EINVAL;
+	struct inode *inode = file_inode(file);
+	struct binderfs_device __user *device = (struct binderfs_device __user *)arg;
+	struct binderfs_device device_req;
+
+	switch (cmd) {
+	case BINDER_CTL_ADD:
+		ret = copy_from_user(&device_req, device, sizeof(device_req));
+		if (ret) {
+			ret = -EFAULT;
+			break;
+		}
+
+		ret = binderfs_binder_device_create(inode, device, &device_req);
+		break;
+	default:
+		break;
+	}
+
+	return ret;
+}
+
+static void binderfs_evict_inode(struct inode *inode)
+{
+	struct binder_device *device = inode->i_private;
+	struct binderfs_info *info = BINDERFS_SB(inode->i_sb);
+
+	clear_inode(inode);
+
+	if (!S_ISCHR(inode->i_mode) || !device)
+		return;
+
+	mutex_lock(&binderfs_minors_mutex);
+	--info->device_count;
+	ida_free(&binderfs_minors, device->minor);
+	mutex_unlock(&binderfs_minors_mutex);
+
+	/* ctx is null for binder-control, but this function ignores null pointers */
+	rust_binder_remove_context(device->ctx);
+
+	kfree(device);
+}
+
+static int binderfs_fs_context_parse_param(struct fs_context *fc,
+					   struct fs_parameter *param)
+{
+	int opt;
+	struct binderfs_mount_opts *ctx = fc->fs_private;
+	struct fs_parse_result result;
+
+	opt = fs_parse(fc, binderfs_fs_parameters, param, &result);
+	if (opt < 0)
+		return opt;
+
+	switch (opt) {
+	case Opt_max:
+		if (result.uint_32 > BINDERFS_MAX_MINOR)
+			return invalfc(fc, "Bad value for '%s'", param->key);
+
+		ctx->max = result.uint_32;
+		break;
+	case Opt_stats_mode:
+		if (!capable(CAP_SYS_ADMIN))
+			return -EPERM;
+
+		ctx->stats_mode = result.uint_32;
+		break;
+	default:
+		return invalfc(fc, "Unsupported parameter '%s'", param->key);
+	}
+
+	return 0;
+}
+
+static int binderfs_fs_context_reconfigure(struct fs_context *fc)
+{
+	struct binderfs_mount_opts *ctx = fc->fs_private;
+	struct binderfs_info *info = BINDERFS_SB(fc->root->d_sb);
+
+	if (info->mount_opts.stats_mode != ctx->stats_mode)
+		return invalfc(fc, "Binderfs stats mode cannot be changed during a remount");
+
+	info->mount_opts.stats_mode = ctx->stats_mode;
+	info->mount_opts.max = ctx->max;
+	return 0;
+}
+
+static int binderfs_show_options(struct seq_file *seq, struct dentry *root)
+{
+	struct binderfs_info *info = BINDERFS_SB(root->d_sb);
+
+	if (info->mount_opts.max <= BINDERFS_MAX_MINOR)
+		seq_printf(seq, ",max=%d", info->mount_opts.max);
+
+	switch (info->mount_opts.stats_mode) {
+	case binderfs_stats_mode_unset:
+		break;
+	case binderfs_stats_mode_global:
+		seq_printf(seq, ",stats=global");
+		break;
+	}
+
+	return 0;
+}
+
+static const struct super_operations binderfs_super_ops = {
+	.evict_inode    = binderfs_evict_inode,
+	.show_options	= binderfs_show_options,
+	.statfs         = simple_statfs,
+};
+
+static inline bool is_binderfs_control_device(const struct dentry *dentry)
+{
+	struct binderfs_info *info = dentry->d_sb->s_fs_info;
+
+	return info->control_dentry == dentry;
+}
+
+static int binderfs_rename(struct mnt_idmap *idmap,
+			   struct inode *old_dir, struct dentry *old_dentry,
+			   struct inode *new_dir, struct dentry *new_dentry,
+			   unsigned int flags)
+{
+	if (is_binderfs_control_device(old_dentry) ||
+	    is_binderfs_control_device(new_dentry))
+		return -EPERM;
+
+	return simple_rename(idmap, old_dir, old_dentry, new_dir,
+			     new_dentry, flags);
+}
+
+static int binderfs_unlink(struct inode *dir, struct dentry *dentry)
+{
+	if (is_binderfs_control_device(dentry))
+		return -EPERM;
+
+	return simple_unlink(dir, dentry);
+}
+
+static const struct file_operations binder_ctl_fops = {
+	.owner		= THIS_MODULE,
+	.open		= nonseekable_open,
+	.unlocked_ioctl	= binder_ctl_ioctl,
+	.compat_ioctl	= binder_ctl_ioctl,
+	.llseek		= noop_llseek,
+};
+
+/**
+ * binderfs_binder_ctl_create - create a new binder-control device
+ * @sb: super block of the binderfs mount
+ *
+ * This function creates a new binder-control device node in the binderfs mount
+ * referred to by @sb.
+ *
+ * Return: 0 on success, negative errno on failure
+ */
+static int binderfs_binder_ctl_create(struct super_block *sb)
+{
+	int minor, ret;
+	struct dentry *dentry;
+	struct binder_device *device;
+	struct inode *inode = NULL;
+	struct dentry *root = sb->s_root;
+	struct binderfs_info *info = sb->s_fs_info;
+#if defined(CONFIG_IPC_NS)
+	bool use_reserve = (info->ipc_ns == &init_ipc_ns);
+#else
+	bool use_reserve = true;
+#endif
+
+	device = kzalloc(sizeof(*device), GFP_KERNEL);
+	if (!device)
+		return -ENOMEM;
+
+	/* If we have already created a binder-control node, return. */
+	if (info->control_dentry) {
+		ret = 0;
+		goto out;
+	}
+
+	ret = -ENOMEM;
+	inode = new_inode(sb);
+	if (!inode)
+		goto out;
+
+	/* Reserve a new minor number for the new device. */
+	mutex_lock(&binderfs_minors_mutex);
+	minor = ida_alloc_max(&binderfs_minors,
+			      use_reserve ? BINDERFS_MAX_MINOR :
+					    BINDERFS_MAX_MINOR_CAPPED,
+			      GFP_KERNEL);
+	mutex_unlock(&binderfs_minors_mutex);
+	if (minor < 0) {
+		ret = minor;
+		goto out;
+	}
+
+	inode->i_ino = SECOND_INODE;
+	simple_inode_init_ts(inode);
+	init_special_inode(inode, S_IFCHR | 0600,
+			   MKDEV(MAJOR(binderfs_dev), minor));
+	inode->i_fop = &binder_ctl_fops;
+	inode->i_uid = info->root_uid;
+	inode->i_gid = info->root_gid;
+
+	device->minor = minor;
+	device->ctx = NULL;
+
+	dentry = d_alloc_name(root, "binder-control");
+	if (!dentry)
+		goto out;
+
+	inode->i_private = device;
+	info->control_dentry = dentry;
+	d_add(dentry, inode);
+
+	return 0;
+
+out:
+	kfree(device);
+	iput(inode);
+
+	return ret;
+}
+
+static const struct inode_operations binderfs_dir_inode_operations = {
+	.lookup = simple_lookup,
+	.rename = binderfs_rename,
+	.unlink = binderfs_unlink,
+};
+
+static struct inode *binderfs_make_inode(struct super_block *sb, int mode)
+{
+	struct inode *ret;
+
+	ret = new_inode(sb);
+	if (ret) {
+		ret->i_ino = iunique(sb, BINDERFS_MAX_MINOR + INODE_OFFSET);
+		ret->i_mode = mode;
+		simple_inode_init_ts(ret);
+	}
+	return ret;
+}
+
+static struct dentry *binderfs_create_dentry(struct dentry *parent,
+					     const char *name)
+{
+	struct dentry *dentry;
+
+	dentry = lookup_one_len(name, parent, strlen(name));
+	if (IS_ERR(dentry))
+		return dentry;
+
+	/* Return error if the file/dir already exists. */
+	if (d_really_is_positive(dentry)) {
+		dput(dentry);
+		return ERR_PTR(-EEXIST);
+	}
+
+	return dentry;
+}
+
+void rust_binderfs_remove_file(struct dentry *dentry)
+{
+	struct inode *parent_inode;
+
+	parent_inode = d_inode(dentry->d_parent);
+	inode_lock(parent_inode);
+	if (simple_positive(dentry)) {
+		dget(dentry);
+		simple_unlink(parent_inode, dentry);
+		d_delete(dentry);
+		dput(dentry);
+	}
+	inode_unlock(parent_inode);
+}
+
+struct dentry *rust_binderfs_create_file(struct dentry *parent, const char *name,
+					 const struct file_operations *fops,
+					 void *data)
+{
+	struct dentry *dentry;
+	struct inode *new_inode, *parent_inode;
+	struct super_block *sb;
+
+	parent_inode = d_inode(parent);
+	inode_lock(parent_inode);
+
+	dentry = binderfs_create_dentry(parent, name);
+	if (IS_ERR(dentry))
+		goto out;
+
+	sb = parent_inode->i_sb;
+	new_inode = binderfs_make_inode(sb, S_IFREG | 0444);
+	if (!new_inode) {
+		dput(dentry);
+		dentry = ERR_PTR(-ENOMEM);
+		goto out;
+	}
+
+	new_inode->i_fop = fops;
+	new_inode->i_private = data;
+	d_instantiate(dentry, new_inode);
+	fsnotify_create(parent_inode, dentry);
+
+out:
+	inode_unlock(parent_inode);
+	return dentry;
+}
+
+struct dentry *rust_binderfs_create_proc_file(struct inode *nodp, int pid)
+{
+	struct binderfs_info *info = nodp->i_sb->s_fs_info;
+	struct dentry *dir = info->proc_log_dir;
+	char strbuf[20 + 1];
+	void *data = (void *)(unsigned long) pid;
+
+	if (!dir)
+		return NULL;
+
+	snprintf(strbuf, sizeof(strbuf), "%u", pid);
+	return rust_binderfs_create_file(dir, strbuf, &rust_binder_proc_fops, data);
+}
+
+static struct dentry *binderfs_create_dir(struct dentry *parent,
+					  const char *name)
+{
+	struct dentry *dentry;
+	struct inode *new_inode, *parent_inode;
+	struct super_block *sb;
+
+	parent_inode = d_inode(parent);
+	inode_lock(parent_inode);
+
+	dentry = binderfs_create_dentry(parent, name);
+	if (IS_ERR(dentry))
+		goto out;
+
+	sb = parent_inode->i_sb;
+	new_inode = binderfs_make_inode(sb, S_IFDIR | 0755);
+	if (!new_inode) {
+		dput(dentry);
+		dentry = ERR_PTR(-ENOMEM);
+		goto out;
+	}
+
+	new_inode->i_fop = &simple_dir_operations;
+	new_inode->i_op = &simple_dir_inode_operations;
+
+	set_nlink(new_inode, 2);
+	d_instantiate(dentry, new_inode);
+	inc_nlink(parent_inode);
+	fsnotify_mkdir(parent_inode, dentry);
+
+out:
+	inode_unlock(parent_inode);
+	return dentry;
+}
+
+static int binder_features_show(struct seq_file *m, void *unused)
+{
+	bool *feature = m->private;
+
+	seq_printf(m, "%d\n", *feature);
+
+	return 0;
+}
+DEFINE_SHOW_ATTRIBUTE(binder_features);
+
+static int init_binder_features(struct super_block *sb)
+{
+	struct dentry *dentry, *dir;
+
+	dir = binderfs_create_dir(sb->s_root, "features");
+	if (IS_ERR(dir))
+		return PTR_ERR(dir);
+
+	dentry = rust_binderfs_create_file(dir, "oneway_spam_detection",
+				      &binder_features_fops,
+				      &binder_features.oneway_spam_detection);
+	if (IS_ERR(dentry))
+		return PTR_ERR(dentry);
+
+	dentry = rust_binderfs_create_file(dir, "extended_error",
+				      &binder_features_fops,
+				      &binder_features.extended_error);
+	if (IS_ERR(dentry))
+		return PTR_ERR(dentry);
+
+	return 0;
+}
+
+static int init_binder_logs(struct super_block *sb)
+{
+	struct dentry *binder_logs_root_dir, *dentry, *proc_log_dir;
+	struct binderfs_info *info;
+	int ret = 0;
+
+	binder_logs_root_dir = binderfs_create_dir(sb->s_root,
+						   "binder_logs");
+	if (IS_ERR(binder_logs_root_dir)) {
+		ret = PTR_ERR(binder_logs_root_dir);
+		goto out;
+	}
+
+	dentry = rust_binderfs_create_file(binder_logs_root_dir, "stats",
+				      &rust_binder_stats_fops, NULL);
+	if (IS_ERR(dentry)) {
+		ret = PTR_ERR(dentry);
+		goto out;
+	}
+
+	dentry = rust_binderfs_create_file(binder_logs_root_dir, "state",
+				      &rust_binder_state_fops, NULL);
+	if (IS_ERR(dentry)) {
+		ret = PTR_ERR(dentry);
+		goto out;
+	}
+
+	dentry = rust_binderfs_create_file(binder_logs_root_dir, "transactions",
+				      &rust_binder_transactions_fops, NULL);
+	if (IS_ERR(dentry)) {
+		ret = PTR_ERR(dentry);
+		goto out;
+	}
+
+	proc_log_dir = binderfs_create_dir(binder_logs_root_dir, "proc");
+	if (IS_ERR(proc_log_dir)) {
+		ret = PTR_ERR(proc_log_dir);
+		goto out;
+	}
+	info = sb->s_fs_info;
+	info->proc_log_dir = proc_log_dir;
+
+out:
+	return ret;
+}
+
+static int binderfs_fill_super(struct super_block *sb, struct fs_context *fc)
+{
+	int ret;
+	struct binderfs_info *info;
+	struct binderfs_mount_opts *ctx = fc->fs_private;
+	struct inode *inode = NULL;
+	struct binderfs_device device_info = {};
+	const char *name;
+	size_t len;
+
+	sb->s_blocksize = PAGE_SIZE;
+	sb->s_blocksize_bits = PAGE_SHIFT;
+
+	/*
+	 * The binderfs filesystem can be mounted by userns root in a
+	 * non-initial userns. By default such mounts have the SB_I_NODEV flag
+	 * set in s_iflags to prevent security issues where userns root can
+	 * just create random device nodes via mknod() since it owns the
+	 * filesystem mount. But binderfs does not allow to create any files
+	 * including devices nodes. The only way to create binder devices nodes
+	 * is through the binder-control device which userns root is explicitly
+	 * allowed to do. So removing the SB_I_NODEV flag from s_iflags is both
+	 * necessary and safe.
+	 */
+	sb->s_iflags &= ~SB_I_NODEV;
+	sb->s_iflags |= SB_I_NOEXEC;
+	sb->s_magic = RUST_BINDERFS_SUPER_MAGIC;
+	sb->s_op = &binderfs_super_ops;
+	sb->s_time_gran = 1;
+
+	sb->s_fs_info = kzalloc(sizeof(struct binderfs_info), GFP_KERNEL);
+	if (!sb->s_fs_info)
+		return -ENOMEM;
+	info = sb->s_fs_info;
+
+	info->ipc_ns = get_ipc_ns(current->nsproxy->ipc_ns);
+
+	info->root_gid = make_kgid(sb->s_user_ns, 0);
+	if (!gid_valid(info->root_gid))
+		info->root_gid = GLOBAL_ROOT_GID;
+	info->root_uid = make_kuid(sb->s_user_ns, 0);
+	if (!uid_valid(info->root_uid))
+		info->root_uid = GLOBAL_ROOT_UID;
+	info->mount_opts.max = ctx->max;
+	info->mount_opts.stats_mode = ctx->stats_mode;
+
+	inode = new_inode(sb);
+	if (!inode)
+		return -ENOMEM;
+
+	inode->i_ino = FIRST_INODE;
+	inode->i_fop = &simple_dir_operations;
+	inode->i_mode = S_IFDIR | 0755;
+	simple_inode_init_ts(inode);
+	inode->i_op = &binderfs_dir_inode_operations;
+	set_nlink(inode, 2);
+
+	sb->s_root = d_make_root(inode);
+	if (!sb->s_root)
+		return -ENOMEM;
+
+	ret = binderfs_binder_ctl_create(sb);
+	if (ret)
+		return ret;
+
+	name = rust_binder_devices_param;
+	for (len = strcspn(name, ","); len > 0; len = strcspn(name, ",")) {
+		strscpy(device_info.name, name, len + 1);
+		ret = binderfs_binder_device_create(inode, NULL, &device_info);
+		if (ret)
+			return ret;
+		name += len;
+		if (*name == ',')
+			name++;
+	}
+
+	ret = init_binder_features(sb);
+	if (ret)
+		return ret;
+
+	if (info->mount_opts.stats_mode == binderfs_stats_mode_global)
+		return init_binder_logs(sb);
+
+	return 0;
+}
+
+static int binderfs_fs_context_get_tree(struct fs_context *fc)
+{
+	return get_tree_nodev(fc, binderfs_fill_super);
+}
+
+static void binderfs_fs_context_free(struct fs_context *fc)
+{
+	struct binderfs_mount_opts *ctx = fc->fs_private;
+
+	kfree(ctx);
+}
+
+static const struct fs_context_operations binderfs_fs_context_ops = {
+	.free		= binderfs_fs_context_free,
+	.get_tree	= binderfs_fs_context_get_tree,
+	.parse_param	= binderfs_fs_context_parse_param,
+	.reconfigure	= binderfs_fs_context_reconfigure,
+};
+
+static int binderfs_init_fs_context(struct fs_context *fc)
+{
+	struct binderfs_mount_opts *ctx;
+
+	ctx = kzalloc(sizeof(struct binderfs_mount_opts), GFP_KERNEL);
+	if (!ctx)
+		return -ENOMEM;
+
+	ctx->max = BINDERFS_MAX_MINOR;
+	ctx->stats_mode = binderfs_stats_mode_unset;
+
+	fc->fs_private = ctx;
+	fc->ops = &binderfs_fs_context_ops;
+
+	return 0;
+}
+
+static void binderfs_kill_super(struct super_block *sb)
+{
+	struct binderfs_info *info = sb->s_fs_info;
+
+	/*
+	 * During inode eviction struct binderfs_info is needed.
+	 * So first wipe the super_block then free struct binderfs_info.
+	 */
+	kill_litter_super(sb);
+
+	if (info && info->ipc_ns)
+		put_ipc_ns(info->ipc_ns);
+
+	kfree(info);
+}
+
+static struct file_system_type binder_fs_type = {
+	.name			= "binder",
+	.init_fs_context	= binderfs_init_fs_context,
+	.parameters		= binderfs_fs_parameters,
+	.kill_sb		= binderfs_kill_super,
+	.fs_flags		= FS_USERNS_MOUNT,
+};
+
+int init_rust_binderfs(void)
+{
+	int ret;
+	const char *name;
+	size_t len;
+
+	/* Verify that the default binderfs device names are valid. */
+	name = rust_binder_devices_param;
+	for (len = strcspn(name, ","); len > 0; len = strcspn(name, ",")) {
+		if (len > BINDERFS_MAX_NAME)
+			return -E2BIG;
+		name += len;
+		if (*name == ',')
+			name++;
+	}
+
+	/* Allocate new major number for binderfs. */
+	ret = alloc_chrdev_region(&binderfs_dev, 0, BINDERFS_MAX_MINOR,
+				  "rust_binder");
+	if (ret)
+		return ret;
+
+	ret = register_filesystem(&binder_fs_type);
+	if (ret) {
+		unregister_chrdev_region(binderfs_dev, BINDERFS_MAX_MINOR);
+		return ret;
+	}
+
+	return ret;
+}

drivers/android/binder/stats.rs

@@ -0,0 +1,88 @@
+// SPDX-License-Identifier: GPL-2.0
+
+// Copyright (C) 2024 Google LLC.
+
+//! Keep track of statistics for binder_logs.
+
+use crate::defs::*;
+use core::sync::atomic::{AtomicU32, Ordering::Relaxed};
+use kernel::{ioctl::_IOC_NR, seq_file::SeqFile, seq_print};
+
+const BC_COUNT: usize = _IOC_NR(BC_REPLY_SG) as usize + 1;
+const BR_COUNT: usize = _IOC_NR(BR_TRANSACTION_PENDING_FROZEN) as usize + 1;
+
+pub(crate) static GLOBAL_STATS: BinderStats = BinderStats::new();
+
+pub(crate) struct BinderStats {
+    bc: [AtomicU32; BC_COUNT],
+    br: [AtomicU32; BR_COUNT],
+}
+
+impl BinderStats {
+    pub(crate) const fn new() -> Self {
+        const ZERO: AtomicU32 = AtomicU32::new(0);
+
+        Self {
+            bc: [ZERO; BC_COUNT],
+            br: [ZERO; BR_COUNT],
+        }
+    }
+
+    pub(crate) fn inc_bc(&self, bc: u32) {
+        let idx = _IOC_NR(bc) as usize;
+        if let Some(bc_ref) = self.bc.get(idx) {
+            bc_ref.fetch_add(1, Relaxed);
+        }
+    }
+
+    pub(crate) fn inc_br(&self, br: u32) {
+        let idx = _IOC_NR(br) as usize;
+        if let Some(br_ref) = self.br.get(idx) {
+            br_ref.fetch_add(1, Relaxed);
+        }
+    }
+
+    pub(crate) fn debug_print(&self, prefix: &str, m: &SeqFile) {
+        for (i, cnt) in self.bc.iter().enumerate() {
+            let cnt = cnt.load(Relaxed);
+            if cnt > 0 {
+                seq_print!(m, "{}{}: {}\n", prefix, command_string(i), cnt);
+            }
+        }
+        for (i, cnt) in self.br.iter().enumerate() {
+            let cnt = cnt.load(Relaxed);
+            if cnt > 0 {
+                seq_print!(m, "{}{}: {}\n", prefix, return_string(i), cnt);
+            }
+        }
+    }
+}
+
+mod strings {
+    use core::str::from_utf8_unchecked;
+    use kernel::str::CStr;
+
+    extern "C" {
+        static binder_command_strings: [*const core::ffi::c_char; super::BC_COUNT];
+        static binder_return_strings: [*const core::ffi::c_char; super::BR_COUNT];
+    }
+
+    pub(super) fn command_string(i: usize) -> &'static str {
+        // SAFETY: Accessing `binder_command_strings` is always safe.
+        let c_str_ptr = unsafe { binder_command_strings[i] };
+        // SAFETY: The `binder_command_strings` array only contains nul-terminated strings.
+        let bytes = unsafe { CStr::from_char_ptr(c_str_ptr) }.as_bytes();
+        // SAFETY: The `binder_command_strings` array only contains strings with ascii-chars.
+        unsafe { from_utf8_unchecked(bytes) }
+    }
+
+    pub(super) fn return_string(i: usize) -> &'static str {
+        // SAFETY: Accessing `binder_return_strings` is always safe.
+        let c_str_ptr = unsafe { binder_return_strings[i] };
+        // SAFETY: The `binder_command_strings` array only contains nul-terminated strings.
+        let bytes = unsafe { CStr::from_char_ptr(c_str_ptr) }.as_bytes();
+        // SAFETY: The `binder_command_strings` array only contains strings with ascii-chars.
+        unsafe { from_utf8_unchecked(bytes) }
+    }
+}
+use strings::{command_string, return_string};

drivers/android/binder/trace.rs

@@ -0,0 +1,109 @@
+// SPDX-License-Identifier: GPL-2.0
+
+// Copyright (C) 2024 Google LLC.
+
+#![allow(unused_variables)]
+#![allow(unused_imports)]
+
+use crate::{defs::BinderTransactionDataSg, node::Node, thread::Thread, transaction::Transaction};
+
+use kernel::error::Result;
+use kernel::task::{Pid, Task};
+use kernel::tracepoint::declare_trace;
+use kernel::uapi::flat_binder_object;
+
+use core::ffi::{c_int, c_uint, c_ulong};
+
+#[inline]
+pub(crate) fn trace_ioctl(cmd: u32, arg: usize) {}
+
+#[inline]
+pub(crate) fn trace_ioctl_done(ret: Result) {}
+
+#[inline]
+pub(crate) fn trace_read_done(ret: Result) {}
+
+#[inline]
+pub(crate) fn trace_write_done(ret: Result) {}
+
+#[inline]
+pub(crate) fn trace_set_priority(thread: &Task, desired_prio: c_int, new_prio: c_int) {}
+
+#[inline]
+pub(crate) fn vh_set_priority(t: &Transaction, task: &Task) {}
+
+#[inline]
+pub(crate) fn vh_restore_priority(task: &Task) {}
+
+#[inline]
+pub(crate) fn trace_wait_for_work(proc_work: bool, transaction_stack: bool, thread_todo: bool) {}
+
+#[inline]
+pub(crate) fn trace_transaction(reply: bool, t: &Transaction) {}
+
+#[inline]
+pub(crate) fn trace_transaction_received(t: &Transaction) {}
+
+#[inline]
+pub(crate) fn trace_transaction_thread_selected(t: &Transaction, th: &Thread) {}
+
+#[inline]
+pub(crate) fn trace_transaction_node_send(
+    t_debug_id: usize,
+    n: &Node,
+    orig: &flat_binder_object,
+    trans: &flat_binder_object,
+) {
+}
+
+#[inline]
+pub(crate) fn trace_transaction_fd_send(t_debug_id: usize, fd: u32, offset: usize) {}
+
+#[inline]
+pub(crate) fn trace_transaction_fd_recv(t_debug_id: usize, fd: u32, offset: usize) {}
+
+#[inline]
+pub(crate) fn trace_transaction_alloc_buf(debug_id: usize, data: &BinderTransactionDataSg) {
+    let data = data as *const BinderTransactionDataSg;
+}
+
+#[inline]
+pub(crate) fn trace_transaction_buffer_release(debug_id: usize) {}
+
+#[inline]
+pub(crate) fn trace_transaction_failed_buffer_release(debug_id: usize) {}
+
+#[inline]
+pub(crate) fn trace_transaction_update_buffer_release(debug_id: usize) {}
+
+#[inline]
+pub(crate) fn trace_update_page_range(pid: Pid, allocate: bool, start: usize, end: usize) {}
+
+macro_rules! define_wrapper_lru_page_class {
+    ($(fn $name:ident;)*) => {$(
+        kernel::macros::paste! {
+            #[inline]
+            pub(crate) fn [< trace_ $name >](pid: Pid, page_index: usize) {
+            }
+        }
+    )*}
+}
+
+define_wrapper_lru_page_class! {
+    fn alloc_lru_start;
+    fn alloc_lru_end;
+    fn free_lru_start;
+    fn free_lru_end;
+    fn alloc_page_start;
+    fn alloc_page_end;
+    fn unmap_user_start;
+    fn unmap_user_end;
+    fn unmap_kernel_start;
+    fn unmap_kernel_end;
+}
+
+#[inline]
+pub(crate) fn trace_command(cmd: u32) {}
+
+#[inline]
+pub(crate) fn trace_return(ret: u32) {}

drivers/android/binder/transaction.rs

@@ -0,0 +1,546 @@
+// SPDX-License-Identifier: GPL-2.0
+
+// Copyright (C) 2024 Google LLC.
+
+use core::sync::atomic::{AtomicBool, Ordering};
+use kernel::{
+    prelude::*,
+    seq_file::SeqFile,
+    seq_print,
+    sync::{Arc, SpinLock},
+    task::Kuid,
+    time::{ktime_ms_delta, Ktime},
+    types::ScopeGuard,
+};
+
+use crate::{
+    allocation::{Allocation, TranslatedFds},
+    defs::*,
+    error::{BinderError, BinderResult},
+    node::{Node, NodeRef},
+    prio::{self, BinderPriority, PriorityState},
+    process::{Process, ProcessInner},
+    ptr_align,
+    thread::{PushWorkRes, Thread},
+    BinderReturnWriter, DArc, DLArc, DTRWrap, DeliverToRead,
+};
+
+#[pin_data(PinnedDrop)]
+pub(crate) struct Transaction {
+    pub(crate) debug_id: usize,
+    target_node: Option<DArc<Node>>,
+    pub(crate) from_parent: Option<DArc<Transaction>>,
+    pub(crate) from: Arc<Thread>,
+    pub(crate) to: Arc<Process>,
+    #[pin]
+    allocation: SpinLock<Option<Allocation>>,
+    is_outstanding: AtomicBool,
+    set_priority_called: AtomicBool,
+    priority: BinderPriority,
+    #[pin]
+    saved_priority: SpinLock<BinderPriority>,
+    code: u32,
+    pub(crate) flags: u32,
+    data_size: usize,
+    offsets_size: usize,
+    data_address: usize,
+    sender_euid: Kuid,
+    txn_security_ctx_off: Option<usize>,
+    pub(crate) oneway_spam_detected: bool,
+    start_time: Ktime,
+}
+
+kernel::list::impl_list_arc_safe! {
+    impl ListArcSafe<0> for Transaction { untracked; }
+}
+
+impl Transaction {
+    pub(crate) fn new(
+        node_ref: NodeRef,
+        from_parent: Option<DArc<Transaction>>,
+        from: &Arc<Thread>,
+        tr: &BinderTransactionDataSg,
+    ) -> BinderResult<DLArc<Self>> {
+        let debug_id = super::next_debug_id();
+        let trd = &tr.transaction_data;
+        let allow_fds = node_ref.node.flags & FLAT_BINDER_FLAG_ACCEPTS_FDS != 0;
+        let txn_security_ctx = node_ref.node.flags & FLAT_BINDER_FLAG_TXN_SECURITY_CTX != 0;
+        let mut txn_security_ctx_off = if txn_security_ctx { Some(0) } else { None };
+        let to = node_ref.node.owner.clone();
+        let mut alloc = match from.copy_transaction_data(
+            to.clone(),
+            tr,
+            debug_id,
+            allow_fds,
+            txn_security_ctx_off.as_mut(),
+        ) {
+            Ok(alloc) => alloc,
+            Err(err) => {
+                if !err.is_dead() {
+                    pr_warn!("Failure in copy_transaction_data: {:?}", err);
+                }
+                return Err(err);
+            }
+        };
+        let oneway_spam_detected = alloc.oneway_spam_detected;
+        if trd.flags & TF_ONE_WAY != 0 {
+            if from_parent.is_some() {
+                pr_warn!("Oneway transaction should not be in a transaction stack.");
+                return Err(EINVAL.into());
+            }
+            alloc.set_info_oneway_node(node_ref.node.clone());
+        }
+        if trd.flags & TF_CLEAR_BUF != 0 {
+            alloc.set_info_clear_on_drop();
+        }
+        let target_node = node_ref.node.clone();
+        alloc.set_info_target_node(node_ref);
+        let data_address = alloc.ptr;
+
+        let priority =
+            if (trd.flags & TF_ONE_WAY == 0) && prio::is_supported_policy(from.task.policy()) {
+                BinderPriority {
+                    sched_policy: from.task.policy(),
+                    prio: from.task.normal_prio(),
+                }
+            } else {
+                from.process.default_priority
+            };
+
+        Ok(DTRWrap::arc_pin_init(pin_init!(Transaction {
+            debug_id,
+            target_node: Some(target_node),
+            from_parent,
+            sender_euid: from.process.cred.euid(),
+            from: from.clone(),
+            to,
+            code: trd.code,
+            flags: trd.flags,
+            data_size: trd.data_size as _,
+            offsets_size: trd.offsets_size as _,
+            data_address,
+            allocation <- kernel::new_spinlock!(Some(alloc.success()), "Transaction::new"),
+            is_outstanding: AtomicBool::new(false),
+            priority,
+            saved_priority <- kernel::new_spinlock!(BinderPriority::default(), "Transaction::saved_priority"),
+            set_priority_called: AtomicBool::new(false),
+            txn_security_ctx_off,
+            oneway_spam_detected,
+            start_time: Ktime::ktime_get(),
+        }))?)
+    }
+
+    pub(crate) fn new_reply(
+        from: &Arc<Thread>,
+        to: Arc<Process>,
+        tr: &BinderTransactionDataSg,
+        allow_fds: bool,
+    ) -> BinderResult<DLArc<Self>> {
+        let debug_id = super::next_debug_id();
+        let trd = &tr.transaction_data;
+        let mut alloc = match from.copy_transaction_data(to.clone(), tr, debug_id, allow_fds, None)
+        {
+            Ok(alloc) => alloc,
+            Err(err) => {
+                pr_warn!("Failure in copy_transaction_data: {:?}", err);
+                return Err(err);
+            }
+        };
+        let oneway_spam_detected = alloc.oneway_spam_detected;
+        if trd.flags & TF_CLEAR_BUF != 0 {
+            alloc.set_info_clear_on_drop();
+        }
+        Ok(DTRWrap::arc_pin_init(pin_init!(Transaction {
+            debug_id,
+            target_node: None,
+            from_parent: None,
+            sender_euid: from.process.task.euid(),
+            from: from.clone(),
+            to,
+            code: trd.code,
+            flags: trd.flags,
+            data_size: trd.data_size as _,
+            offsets_size: trd.offsets_size as _,
+            data_address: alloc.ptr,
+            allocation <- kernel::new_spinlock!(Some(alloc.success()), "Transaction::new"),
+            is_outstanding: AtomicBool::new(false),
+            priority: BinderPriority::default(),
+            saved_priority <- kernel::new_spinlock!(BinderPriority::default(), "Transaction::saved_priority"),
+            set_priority_called: AtomicBool::new(false),
+            txn_security_ctx_off: None,
+            oneway_spam_detected,
+            start_time: Ktime::ktime_get(),
+        }))?)
+    }
+
+    #[inline(never)]
+    pub(crate) fn debug_print_inner(&self, m: &SeqFile, prefix: &str) {
+        seq_print!(
+            m,
+            "{}{}: from {}:{} to {} code {:x} flags {:x} pri {}:{} elapsed {}ms",
+            prefix,
+            self.debug_id,
+            self.from.process.task.pid(),
+            self.from.id,
+            self.to.task.pid(),
+            self.code,
+            self.flags,
+            self.priority.sched_policy,
+            self.priority.prio,
+            ktime_ms_delta(Ktime::ktime_get(), self.start_time),
+        );
+        if let Some(target_node) = &self.target_node {
+            seq_print!(m, " node {}", target_node.debug_id);
+        }
+        seq_print!(m, " size {}:{}\n", self.data_size, self.offsets_size);
+    }
+
+    pub(crate) fn saved_priority(&self) -> BinderPriority {
+        *self.saved_priority.lock()
+    }
+
+    /// Determines if the transaction is stacked on top of the given transaction.
+    pub(crate) fn is_stacked_on(&self, onext: &Option<DArc<Self>>) -> bool {
+        match (&self.from_parent, onext) {
+            (None, None) => true,
+            (Some(from_parent), Some(next)) => Arc::ptr_eq(from_parent, next),
+            _ => false,
+        }
+    }
+
+    /// Returns a pointer to the next transaction on the transaction stack, if there is one.
+    pub(crate) fn clone_next(&self) -> Option<DArc<Self>> {
+        Some(self.from_parent.as_ref()?.clone())
+    }
+
+    /// Searches in the transaction stack for a thread that belongs to the target process. This is
+    /// useful when finding a target for a new transaction: if the node belongs to a process that
+    /// is already part of the transaction stack, we reuse the thread.
+    fn find_target_thread(&self) -> Option<Arc<Thread>> {
+        let mut it = &self.from_parent;
+        while let Some(transaction) = it {
+            if Arc::ptr_eq(&transaction.from.process, &self.to) {
+                return Some(transaction.from.clone());
+            }
+            it = &transaction.from_parent;
+        }
+        None
+    }
+
+    /// Searches in the transaction stack for a transaction originating at the given thread.
+    pub(crate) fn find_from(&self, thread: &Thread) -> Option<DArc<Transaction>> {
+        let mut it = &self.from_parent;
+        while let Some(transaction) = it {
+            if core::ptr::eq(thread, transaction.from.as_ref()) {
+                return Some(transaction.clone());
+            }
+
+            it = &transaction.from_parent;
+        }
+        None
+    }
+
+    pub(crate) fn set_outstanding(&self, to_process: &mut ProcessInner) {
+        // No race because this method is only called once.
+        if !self.is_outstanding.load(Ordering::Relaxed) {
+            self.is_outstanding.store(true, Ordering::Relaxed);
+            to_process.add_outstanding_txn();
+        }
+    }
+
+    /// Decrement `outstanding_txns` in `to` if it hasn't already been decremented.
+    fn drop_outstanding_txn(&self) {
+        // No race because this is called at most twice, and one of the calls are in the
+        // destructor, which is guaranteed to not race with any other operations on the
+        // transaction. It also cannot race with `set_outstanding`, since submission happens
+        // before delivery.
+        if self.is_outstanding.load(Ordering::Relaxed) {
+            self.is_outstanding.store(false, Ordering::Relaxed);
+            self.to.drop_outstanding_txn();
+        }
+    }
+
+    /// Submits the transaction to a work queue. Uses a thread if there is one in the transaction
+    /// stack, otherwise uses the destination process.
+    ///
+    /// Not used for replies.
+    pub(crate) fn submit(self: DLArc<Self>) -> BinderResult {
+        crate::trace::trace_transaction(false, &self);
+
+        // Defined before `process_inner` so that the destructor runs after releasing the lock.
+        let mut _t_outdated;
+
+        let oneway = self.flags & TF_ONE_WAY != 0;
+        let process = self.to.clone();
+        let mut process_inner = process.inner.lock();
+
+        self.set_outstanding(&mut process_inner);
+
+        if oneway {
+            if let Some(target_node) = self.target_node.clone() {
+                if process_inner.is_frozen {
+                    process_inner.async_recv = true;
+                    if self.flags & TF_UPDATE_TXN != 0 {
+                        if let Some(t_outdated) =
+                            target_node.take_outdated_transaction(&self, &mut process_inner)
+                        {
+                            crate::trace::trace_transaction_update_buffer_release(
+                                t_outdated.debug_id,
+                            );
+                            // Save the transaction to be dropped after locks are released.
+                            _t_outdated = t_outdated;
+                        }
+                    }
+                }
+                match target_node.submit_oneway(self, &mut process_inner) {
+                    Ok(()) => {}
+                    Err((err, work)) => {
+                        drop(process_inner);
+                        // Drop work after releasing process lock.
+                        drop(work);
+                        return Err(err);
+                    }
+                }
+
+                if process_inner.is_frozen {
+                    return Err(BinderError::new_frozen_oneway());
+                } else {
+                    return Ok(());
+                }
+            } else {
+                pr_err!("Failed to submit oneway transaction to node.");
+            }
+        }
+
+        if process_inner.is_frozen {
+            process_inner.sync_recv = true;
+            return Err(BinderError::new_frozen());
+        }
+
+        let res = if let Some(thread) = self.find_target_thread() {
+            match thread.push_work(self) {
+                PushWorkRes::Ok => Ok(()),
+                PushWorkRes::FailedDead(me) => Err((BinderError::new_dead(), me)),
+            }
+        } else {
+            process_inner.push_work(self)
+        };
+        drop(process_inner);
+
+        match res {
+            Ok(()) => Ok(()),
+            Err((err, work)) => {
+                // Drop work after releasing process lock.
+                drop(work);
+                Err(err)
+            }
+        }
+    }
+
+    /// Check whether one oneway transaction can supersede another.
+    pub(crate) fn can_replace(&self, old: &Transaction) -> bool {
+        if self.from.process.task.pid() != old.from.process.task.pid() {
+            return false;
+        }
+
+        if self.flags & old.flags & (TF_ONE_WAY | TF_UPDATE_TXN) != (TF_ONE_WAY | TF_UPDATE_TXN) {
+            return false;
+        }
+
+        let target_node_match = match (self.target_node.as_ref(), old.target_node.as_ref()) {
+            (None, None) => true,
+            (Some(tn1), Some(tn2)) => Arc::ptr_eq(tn1, tn2),
+            _ => false,
+        };
+
+        self.code == old.code && self.flags == old.flags && target_node_match
+    }
+
+    fn prepare_file_list(&self) -> Result<TranslatedFds> {
+        let mut alloc = self.allocation.lock().take().ok_or(ESRCH)?;
+
+        match alloc.translate_fds() {
+            Ok(translated) => {
+                *self.allocation.lock() = Some(alloc);
+                Ok(translated)
+            }
+            Err(err) => {
+                // Free the allocation eagerly.
+                drop(alloc);
+                Err(err)
+            }
+        }
+    }
+}
+
+impl DeliverToRead for Transaction {
+    fn do_work(
+        self: DArc<Self>,
+        thread: &Thread,
+        writer: &mut BinderReturnWriter<'_>,
+    ) -> Result<bool> {
+        let send_failed_reply = ScopeGuard::new(|| {
+            if self.target_node.is_some() && self.flags & TF_ONE_WAY == 0 {
+                let reply = Err(BR_FAILED_REPLY);
+                self.from.deliver_reply(reply, &self);
+            }
+            self.drop_outstanding_txn();
+        });
+
+        // Update thread priority. This only has an effect if the transaction is delivered via the
+        // process work list, since the priority has otherwise already been updated.
+        self.on_thread_selected(thread);
+
+        let files = if let Ok(list) = self.prepare_file_list() {
+            list
+        } else {
+            // On failure to process the list, we send a reply back to the sender and ignore the
+            // transaction on the recipient.
+            return Ok(true);
+        };
+
+        let mut tr_sec = BinderTransactionDataSecctx::default();
+        let tr = tr_sec.tr_data();
+        if let Some(target_node) = &self.target_node {
+            let (ptr, cookie) = target_node.get_id();
+            tr.target.ptr = ptr as _;
+            tr.cookie = cookie as _;
+        };
+        tr.code = self.code;
+        tr.flags = self.flags;
+        tr.data_size = self.data_size as _;
+        tr.data.ptr.buffer = self.data_address as _;
+        tr.offsets_size = self.offsets_size as _;
+        if tr.offsets_size > 0 {
+            tr.data.ptr.offsets = (self.data_address + ptr_align(self.data_size)) as _;
+        }
+        tr.sender_euid = self.sender_euid.into_uid_in_current_ns();
+        tr.sender_pid = 0;
+        if self.target_node.is_some() && self.flags & TF_ONE_WAY == 0 {
+            // Not a reply and not one-way.
+            tr.sender_pid = self.from.process.task.pid_in_current_ns();
+        }
+        let code = if self.target_node.is_none() {
+            BR_REPLY
+        } else if self.txn_security_ctx_off.is_some() {
+            BR_TRANSACTION_SEC_CTX
+        } else {
+            BR_TRANSACTION
+        };
+
+        // Write the transaction code and data to the user buffer.
+        writer.write_code(code)?;
+        if let Some(off) = self.txn_security_ctx_off {
+            tr_sec.secctx = (self.data_address + off) as u64;
+            writer.write_payload(&tr_sec)?;
+        } else {
+            writer.write_payload(&*tr)?;
+        }
+
+        let mut alloc = self.allocation.lock().take().ok_or(ESRCH)?;
+
+        // Dismiss the completion of transaction with a failure. No failure paths are allowed from
+        // here on out.
+        send_failed_reply.dismiss();
+
+        // Commit files, and set FDs in FDA to be closed on buffer free.
+        let close_on_free = files.commit();
+        alloc.set_info_close_on_free(close_on_free);
+
+        // It is now the user's responsibility to clear the allocation.
+        alloc.keep_alive();
+
+        self.drop_outstanding_txn();
+
+        crate::trace::trace_transaction_received(&self);
+
+        // When this is not a reply and not a oneway transaction, update `current_transaction`. If
+        // it's a reply, `current_transaction` has already been updated appropriately.
+        if self.target_node.is_some() && tr_sec.transaction_data.flags & TF_ONE_WAY == 0 {
+            thread.set_current_transaction(self);
+        }
+
+        Ok(false)
+    }
+
+    fn cancel(self: DArc<Self>) {
+        let allocation = self.allocation.lock().take();
+        drop(allocation);
+
+        // If this is not a reply or oneway transaction, then send a dead reply.
+        if self.target_node.is_some() && self.flags & TF_ONE_WAY == 0 {
+            let reply = Err(BR_DEAD_REPLY);
+            self.from.deliver_reply(reply, &self);
+        }
+
+        self.drop_outstanding_txn();
+    }
+
+    fn on_thread_selected(&self, to_thread: &Thread) {
+        // Return immediately if reply.
+        let target_node = match self.target_node.as_ref() {
+            Some(target_node) => target_node,
+            None => return,
+        };
+
+        // We only need to do this once.
+        if self.set_priority_called.swap(true, Ordering::Relaxed) {
+            return;
+        }
+
+        crate::trace::trace_transaction_thread_selected(self, to_thread);
+
+        let node_prio = target_node.node_prio();
+        let mut desired = self.priority;
+
+        if !target_node.inherit_rt() && prio::is_rt_policy(desired.sched_policy) {
+            desired.prio = prio::DEFAULT_PRIO;
+            desired.sched_policy = prio::SCHED_NORMAL;
+        }
+
+        if node_prio.prio < self.priority.prio
+            || (node_prio.prio == self.priority.prio && node_prio.sched_policy == prio::SCHED_FIFO)
+        {
+            // In case the minimum priority on the node is
+            // higher (lower value), use that priority. If
+            // the priority is the same, but the node uses
+            // SCHED_FIFO, prefer SCHED_FIFO, since it can
+            // run unbounded, unlike SCHED_RR.
+            desired = node_prio;
+        }
+
+        let mut prio_state = to_thread.prio_lock.lock();
+        if prio_state.state == PriorityState::Pending {
+            // Task is in the process of changing priorities
+            // saving its current values would be incorrect.
+            // Instead, save the pending priority and signal
+            // the task to abort the priority restore.
+            prio_state.state = PriorityState::Abort;
+            *self.saved_priority.lock() = prio_state.next;
+        } else {
+            let task = &*self.to.task;
+            let mut saved_priority = self.saved_priority.lock();
+            saved_priority.sched_policy = task.policy();
+            saved_priority.prio = task.normal_prio();
+        }
+        drop(prio_state);
+
+        to_thread.set_priority(&desired, self);
+    }
+
+    fn should_sync_wakeup(&self) -> bool {
+        self.flags & TF_ONE_WAY == 0
+    }
+
+    fn debug_print(&self, m: &SeqFile, _prefix: &str, tprefix: &str) -> Result<()> {
+        self.debug_print_inner(m, tprefix);
+        Ok(())
+    }
+}
+
+#[pinned_drop]
+impl PinnedDrop for Transaction {
+    fn drop(self: Pin<&mut Self>) {
+        self.drop_outstanding_txn();
+    }
+}