2022-07-25 23:46:47 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
rust: workaround `bindgen` issue with forward references to `enum` types
`bindgen` currently generates the wrong type for an `enum` when there
is a forward reference to it. For instance:
enum E;
enum E { A };
generates:
pub const E_A: E = 0;
pub type E = i32;
instead of the expected:
pub const E_A: E = 0;
pub type E = ffi::c_uint;
The issue was reported to upstream `bindgen` [1].
Now, both GCC and Clang support silently these forward references to
`enum` types, unless `-Wpedantic` is passed, and it turns out that some
headers in the kernel depend on them.
Thus, depending on how the headers are included, which in turn may depend
on the kernel configuration or the architecture, we may get a different
type on the Rust side for a given C `enum`.
That can be quite confusing, to say the least, especially since
developers may only notice issues when building for other architectures
like in [2]. In particular, they may end up forcing a cast and adding
an `#[allow(clippy::unnecessary_cast)]` like it was done in commit
94e05a66ea3e ("rust: hrtimer: allow timer restart from timer handler"),
which isn't great.
Instead, let's have a section at the top of our `bindings_helper.h` that
`#include`s the headers with the affected types -- hopefully there are
not many cases and there is a single ordering that covers all cases.
This allows us to remove the cast and the `#[allow]`, thus keeping the
correct code in the source files. When the issue gets resolved in upstream
`bindgen` (and we update our minimum `bindgen` version), we can easily
remove this section at the top.
Link: https://github.com/rust-lang/rust-bindgen/issues/3179 [1]
Link: https://lore.kernel.org/rust-for-linux/87tt7md1s6.fsf@kernel.org/ [2]
Acked-by: Andreas Hindborg <a.hindborg@kernel.org>
Link: https://lore.kernel.org/r/20250325184309.97170-1-ojeda@kernel.org
[ Added extra paragraph on the comment to clarify that the workaround may
not be possible in some cases. - Miguel ]
Signed-off-by: Miguel Ojeda <ojeda@kernel.org>
2025-03-25 19:43:09 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2025-06-20 16:39:13 +01:00
|
|
|
|
2025-04-11 01:55:23 +02:00
|
|
|
|
2025-04-11 01:55:22 +02:00
|
|
|
|
2025-04-11 01:55:25 +02:00
|
|
|
|
2025-04-11 01:55:26 +02:00
|
|
|
|
2025-04-11 01:55:21 +02:00
|
|
|
|
rust: support running Rust documentation tests as KUnit ones
Rust has documentation tests: these are typically examples of
usage of any item (e.g. function, struct, module...).
They are very convenient because they are just written
alongside the documentation. For instance:
/// Sums two numbers.
///
/// ```
/// assert_eq!(mymod::f(10, 20), 30);
/// ```
pub fn f(a: i32, b: i32) -> i32 {
a + b
}
In userspace, the tests are collected and run via `rustdoc`.
Using the tool as-is would be useful already, since it allows
to compile-test most tests (thus enforcing they are kept
in sync with the code they document) and run those that do not
depend on in-kernel APIs.
However, by transforming the tests into a KUnit test suite,
they can also be run inside the kernel. Moreover, the tests
get to be compiled as other Rust kernel objects instead of
targeting userspace.
On top of that, the integration with KUnit means the Rust
support gets to reuse the existing testing facilities. For
instance, the kernel log would look like:
KTAP version 1
1..1
KTAP version 1
# Subtest: rust_doctests_kernel
1..59
# rust_doctest_kernel_build_assert_rs_0.location: rust/kernel/build_assert.rs:13
ok 1 rust_doctest_kernel_build_assert_rs_0
# rust_doctest_kernel_build_assert_rs_1.location: rust/kernel/build_assert.rs:56
ok 2 rust_doctest_kernel_build_assert_rs_1
# rust_doctest_kernel_init_rs_0.location: rust/kernel/init.rs:122
ok 3 rust_doctest_kernel_init_rs_0
...
# rust_doctest_kernel_types_rs_2.location: rust/kernel/types.rs:150
ok 59 rust_doctest_kernel_types_rs_2
# rust_doctests_kernel: pass:59 fail:0 skip:0 total:59
# Totals: pass:59 fail:0 skip:0 total:59
ok 1 rust_doctests_kernel
Therefore, add support for running Rust documentation tests
in KUnit. Some other notes about the current implementation
and support follow.
The transformation is performed by a couple scripts written
as Rust hostprogs.
Tests using the `?` operator are also supported as usual, e.g.:
/// ```
/// # use kernel::{spawn_work_item, workqueue};
/// spawn_work_item!(workqueue::system(), || pr_info!("x"))?;
/// # Ok::<(), Error>(())
/// ```
The tests are also compiled with Clippy under `CLIPPY=1`, just
like normal code, thus also benefitting from extra linting.
The names of the tests are currently automatically generated.
This allows to reduce the burden for documentation writers,
while keeping them fairly stable for bisection. This is an
improvement over the `rustdoc`-generated names, which include
the line number; but ideally we would like to get `rustdoc` to
provide the Rust item path and a number (for multiple examples
in a single documented Rust item).
In order for developers to easily see from which original line
a failed doctests came from, a KTAP diagnostic line is printed
to the log, containing the location (file and line) of the
original test (i.e. instead of the location in the generated
Rust file):
# rust_doctest_kernel_types_rs_2.location: rust/kernel/types.rs:150
This line follows the syntax for declaring test metadata in the
proposed KTAP v2 spec [1], which may be used for the proposed
KUnit test attributes API [2]. Thus hopefully this will make
migration easier later on (suggested by David [3]).
The original line in that test attribute is figured out by
providing an anchor (suggested by Boqun [4]). The original file
is found by walking the filesystem, checking directory prefixes
to reduce the amount of combinations to check, and it is only
done once per file. Ambiguities are detected and reported.
A notable difference from KUnit C tests is that the Rust tests
appear to assert using the usual `assert!` and `assert_eq!`
macros from the Rust standard library (`core`). We provide
a custom version that forwards the call to KUnit instead.
Importantly, these macros do not require passing context,
unlike the KUnit C ones (i.e. `struct kunit *`). This makes
them easier to use, and readers of the documentation do not need
to care about which testing framework is used. In addition, it
may allow us to test third-party code more easily in the future.
However, a current limitation is that KUnit does not support
assertions in other tasks. Thus we presently simply print an
error to the kernel log if an assertion actually failed. This
should be revisited to properly fail the test, perhaps saving
the context somewhere else, or letting KUnit handle it.
Link: https://lore.kernel.org/lkml/20230420205734.1288498-1-rmoar@google.com/ [1]
Link: https://lore.kernel.org/linux-kselftest/20230707210947.1208717-1-rmoar@google.com/ [2]
Link: https://lore.kernel.org/rust-for-linux/CABVgOSkOLO-8v6kdAGpmYnZUb+LKOX0CtYCo-Bge7r_2YTuXDQ@mail.gmail.com/ [3]
Link: https://lore.kernel.org/rust-for-linux/ZIps86MbJF%2FiGIzd@boqun-archlinux/ [4]
Signed-off-by: Miguel Ojeda <ojeda@kernel.org>
Reviewed-by: David Gow <davidgow@google.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
2023-07-18 07:27:51 +02:00
|
|
|
|
2025-04-14 15:18:06 +02:00
|
|
|
|
2025-09-08 07:21:51 +00:00
|
|
|
|
2024-06-11 13:45:49 +02:00
|
|
|
|
2024-08-09 13:28:35 +00:00
|
|
|
|
2024-06-28 11:11:52 +02:00
|
|
|
|
2025-02-20 16:34:00 +05:30
|
|
|
|
2025-06-12 14:17:13 +02:00
|
|
|
|
2025-05-08 10:52:09 +02:00
|
|
|
|
2025-01-08 16:18:38 +05:30
|
|
|
|
2025-01-07 12:51:08 +05:30
|
|
|
|
2025-02-24 18:39:35 -05:00
|
|
|
|
2024-09-15 14:31:30 +00:00
|
|
|
|
2025-09-04 21:13:52 +00:00
|
|
|
|
2025-02-10 13:30:26 +01:00
|
|
|
|
2025-08-28 15:32:14 +02:00
|
|
|
|
2025-03-17 20:52:09 +02:00
|
|
|
|
2023-05-31 17:44:50 +00:00
|
|
|
|
2023-12-13 09:42:08 +09:00
|
|
|
|
rust_binder: add Rust Binder driver
We're generally not proponents of rewrites (nasty uncomfortable things
that make you late for dinner!). So why rewrite Binder?
Binder has been evolving over the past 15+ years to meet the evolving
needs of Android. Its responsibilities, expectations, and complexity
have grown considerably during that time. While we expect Binder to
continue to evolve along with Android, there are a number of factors
that currently constrain our ability to develop/maintain it. Briefly
those are:
1. Complexity: Binder is at the intersection of everything in Android and
fulfills many responsibilities beyond IPC. It has become many things
to many people, and due to its many features and their interactions
with each other, its complexity is quite high. In just 6kLOC it must
deliver transactions to the right threads. It must correctly parse
and translate the contents of transactions, which can contain several
objects of different types (e.g., pointers, fds) that can interact
with each other. It controls the size of thread pools in userspace,
and ensures that transactions are assigned to threads in ways that
avoid deadlocks where the threadpool has run out of threads. It must
track refcounts of objects that are shared by several processes by
forwarding refcount changes between the processes correctly. It must
handle numerous error scenarios and it combines/nests 13 different
locks, 7 reference counters, and atomic variables. Finally, It must
do all of this as fast and efficiently as possible. Minor performance
regressions can cause a noticeably degraded user experience.
2. Things to improve: Thousand-line functions [1], error-prone error
handling [2], and confusing structure can occur as a code base grows
organically. After more than a decade of development, this codebase
could use an overhaul.
[1]: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/android/binder.c?h=v6.5#n2896
[2]: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/android/binder.c?h=v6.5#n3658
3. Security critical: Binder is a critical part of Android's sandboxing
strategy. Even Android's most de-privileged sandboxes (e.g. the
Chrome renderer, or SW Codec) have direct access to Binder. More than
just about any other component, it's important that Binder provide
robust security, and itself be robust against security
vulnerabilities.
It's #1 (high complexity) that has made continuing to evolve Binder and
resolving #2 (tech debt) exceptionally difficult without causing #3
(security issues). For Binder to continue to meet Android's needs, we
need better ways to manage (and reduce!) complexity without increasing
the risk.
The biggest change is obviously the choice of programming language. We
decided to use Rust because it directly addresses a number of the
challenges within Binder that we have faced during the last years. It
prevents mistakes with ref counting, locking, bounds checking, and also
does a lot to reduce the complexity of error handling. Additionally,
we've been able to use the more expressive type system to encode the
ownership semantics of the various structs and pointers, which takes the
complexity of managing object lifetimes out of the hands of the
programmer, reducing the risk of use-after-frees and similar problems.
Rust has many different pointer types that it uses to encode ownership
semantics into the type system, and this is probably one of the most
important aspects of how it helps in Binder. The Binder driver has a lot
of different objects that have complex ownership semantics; some
pointers own a refcount, some pointers have exclusive ownership, and
some pointers just reference the object and it is kept alive in some
other manner. With Rust, we can use a different pointer type for each
kind of pointer, which enables the compiler to enforce that the
ownership semantics are implemented correctly.
Another useful feature is Rust's error handling. Rust allows for more
simplified error handling with features such as destructors, and you get
compilation failures if errors are not properly handled. This means that
even though Rust requires you to spend more lines of code than C on
things such as writing down invariants that are left implicit in C, the
Rust driver is still slightly smaller than C binder: Rust is 5.5kLOC and
C is 5.8kLOC. (These numbers are excluding blank lines, comments,
binderfs, and any debugging facilities in C that are not yet implemented
in the Rust driver. The numbers include abstractions in rust/kernel/
that are unlikely to be used by other drivers than Binder.)
Although this rewrite completely rethinks how the code is structured and
how assumptions are enforced, we do not fundamentally change *how* the
driver does the things it does. A lot of careful thought has gone into
the existing design. The rewrite is aimed rather at improving code
health, structure, readability, robustness, security, maintainability
and extensibility. We also include more inline documentation, and
improve how assumptions in the code are enforced. Furthermore, all
unsafe code is annotated with a SAFETY comment that explains why it is
correct.
We have left the binderfs filesystem component in C. Rewriting it in
Rust would be a large amount of work and requires a lot of bindings to
the file system interfaces. Binderfs has not historically had the same
challenges with security and complexity, so rewriting binderfs seems to
have lower value than the rest of Binder.
Correctness and feature parity
------------------------------
Rust binder passes all tests that validate the correctness of Binder in
the Android Open Source Project. We can boot a device, and run a variety
of apps and functionality without issues. We have performed this both on
the Cuttlefish Android emulator device, and on a Pixel 6 Pro.
As for feature parity, Rust binder currently implements all features
that C binder supports, with the exception of some debugging facilities.
The missing debugging facilities will be added before we submit the Rust
implementation upstream.
Tracepoints
-----------
I did not include all of the tracepoints as I felt that the mechansim
for making C access fields of Rust structs should be discussed on list
separately. I also did not include the support for building Rust Binder
as a module since that requires exporting a bunch of additional symbols
on the C side.
Original RFC Link with old benchmark numbers:
https://lore.kernel.org/r/20231101-rust-binder-v1-0-08ba9197f637@google.com
Co-developed-by: Wedson Almeida Filho <wedsonaf@gmail.com>
Signed-off-by: Wedson Almeida Filho <wedsonaf@gmail.com>
Co-developed-by: Matt Gilbride <mattgilbride@google.com>
Signed-off-by: Matt Gilbride <mattgilbride@google.com>
Acked-by: Carlos Llamas <cmllamas@google.com>
Acked-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: Alice Ryhl <aliceryhl@google.com>
Link: https://lore.kernel.org/r/20250919-rust-binder-v2-1-a384b09f28dd@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2025-09-19 06:42:07 +00:00
|
|
|
|
rust: file: add Rust abstraction for `struct file`
This abstraction makes it possible to manipulate the open files for a
process. The new `File` struct wraps the C `struct file`. When accessing
it using the smart pointer `ARef<File>`, the pointer will own a
reference count to the file. When accessing it as `&File`, then the
reference does not own a refcount, but the borrow checker will ensure
that the reference count does not hit zero while the `&File` is live.
Since this is intended to manipulate the open files of a process, we
introduce an `fget` constructor that corresponds to the C `fget`
method. In future patches, it will become possible to create a new fd in
a process and bind it to a `File`. Rust Binder will use these to send
fds from one process to another.
We also provide a method for accessing the file's flags. Rust Binder
will use this to access the flags of the Binder fd to check whether the
non-blocking flag is set, which affects what the Binder ioctl does.
This introduces a struct for the EBADF error type, rather than just
using the Error type directly. This has two advantages:
* `File::fget` returns a `Result<ARef<File>, BadFdError>`, which the
compiler will represent as a single pointer, with null being an error.
This is possible because the compiler understands that `BadFdError`
has only one possible value, and it also understands that the
`ARef<File>` smart pointer is guaranteed non-null.
* Additionally, we promise to users of the method that the method can
only fail with EBADF, which means that they can rely on this promise
without having to inspect its implementation.
That said, there are also two disadvantages:
* Defining additional error types involves boilerplate.
* The question mark operator will only utilize the `From` trait once,
which prevents you from using the question mark operator on
`BadFdError` in methods that return some third error type that the
kernel `Error` is convertible into. (However, it works fine in methods
that return `Error`.)
Signed-off-by: Wedson Almeida Filho <wedsonaf@gmail.com>
Co-developed-by: Daniel Xu <dxu@dxuuu.xyz>
Signed-off-by: Daniel Xu <dxu@dxuuu.xyz>
Co-developed-by: Alice Ryhl <aliceryhl@google.com>
Reviewed-by: Benno Lossin <benno.lossin@proton.me>
Signed-off-by: Alice Ryhl <aliceryhl@google.com>
Link: https://lore.kernel.org/r/20240915-alice-file-v10-3-88484f7a3dcf@google.com
Reviewed-by: Gary Guo <gary@garyguo.net>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-09-15 14:31:29 +00:00
|
|
|
|
2024-06-18 17:48:35 +02:00
|
|
|
|
rust: file: add Rust abstraction for `struct file`
This abstraction makes it possible to manipulate the open files for a
process. The new `File` struct wraps the C `struct file`. When accessing
it using the smart pointer `ARef<File>`, the pointer will own a
reference count to the file. When accessing it as `&File`, then the
reference does not own a refcount, but the borrow checker will ensure
that the reference count does not hit zero while the `&File` is live.
Since this is intended to manipulate the open files of a process, we
introduce an `fget` constructor that corresponds to the C `fget`
method. In future patches, it will become possible to create a new fd in
a process and bind it to a `File`. Rust Binder will use these to send
fds from one process to another.
We also provide a method for accessing the file's flags. Rust Binder
will use this to access the flags of the Binder fd to check whether the
non-blocking flag is set, which affects what the Binder ioctl does.
This introduces a struct for the EBADF error type, rather than just
using the Error type directly. This has two advantages:
* `File::fget` returns a `Result<ARef<File>, BadFdError>`, which the
compiler will represent as a single pointer, with null being an error.
This is possible because the compiler understands that `BadFdError`
has only one possible value, and it also understands that the
`ARef<File>` smart pointer is guaranteed non-null.
* Additionally, we promise to users of the method that the method can
only fail with EBADF, which means that they can rely on this promise
without having to inspect its implementation.
That said, there are also two disadvantages:
* Defining additional error types involves boilerplate.
* The question mark operator will only utilize the `From` trait once,
which prevents you from using the question mark operator on
`BadFdError` in methods that return some third error type that the
kernel `Error` is convertible into. (However, it works fine in methods
that return `Error`.)
Signed-off-by: Wedson Almeida Filho <wedsonaf@gmail.com>
Co-developed-by: Daniel Xu <dxu@dxuuu.xyz>
Signed-off-by: Daniel Xu <dxu@dxuuu.xyz>
Co-developed-by: Alice Ryhl <aliceryhl@google.com>
Reviewed-by: Benno Lossin <benno.lossin@proton.me>
Signed-off-by: Alice Ryhl <aliceryhl@google.com>
Link: https://lore.kernel.org/r/20240915-alice-file-v10-3-88484f7a3dcf@google.com
Reviewed-by: Gary Guo <gary@garyguo.net>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-09-15 14:31:29 +00:00
|
|
|
|
2025-11-16 16:21:44 +00:00
|
|
|
|
2026-01-18 10:08:08 +00:00
|
|
|
|
|
|
|
|
|
2025-07-17 12:55:22 -03:00
|
|
|
|
2024-01-08 14:49:58 +00:00
|
|
|
|
2024-10-30 16:04:24 +00:00
|
|
|
|
2023-12-13 09:42:08 +09:00
|
|
|
|
2025-08-20 16:53:37 +02:00
|
|
|
|
2024-10-01 08:22:22 +00:00
|
|
|
|
2024-12-19 18:04:15 +01:00
|
|
|
|
2024-12-19 18:04:11 +01:00
|
|
|
|
2023-12-13 09:42:08 +09:00
|
|
|
|
rust: file: add `Kuid` wrapper
Adds a wrapper around `kuid_t` called `Kuid`. This allows us to define
various operations on kuids such as equality and current_euid. It also
lets us provide conversions from kuid into userspace values.
Rust Binder needs these operations because it needs to compare kuids for
equality, and it needs to tell userspace about the pid and uid of
incoming transactions.
To read kuids from a `struct task_struct`, you must currently use
various #defines that perform the appropriate field access under an RCU
read lock. Currently, we do not have a Rust wrapper for rcu_read_lock,
which means that for this patch, there are two ways forward:
1. Inline the methods into Rust code, and use __rcu_read_lock directly
rather than the rcu_read_lock wrapper. This gives up lockdep for
these usages of RCU.
2. Wrap the various #defines in helpers and call the helpers from Rust.
This patch uses the second option. One possible disadvantage of the
second option is the possible introduction of speculation gadgets, but
as discussed in [1], the risk appears to be acceptable.
Of course, once a wrapper for rcu_read_lock is available, it is
preferable to use that over either of the two above approaches.
Link: https://lore.kernel.org/all/202312080947.674CD2DC7@keescook/ [1]
Reviewed-by: Benno Lossin <benno.lossin@proton.me>
Reviewed-by: Martin Rodriguez Reboredo <yakoyoku@gmail.com>
Reviewed-by: Trevor Gross <tmgross@umich.edu>
Signed-off-by: Alice Ryhl <aliceryhl@google.com>
Link: https://lore.kernel.org/r/20240915-alice-file-v10-7-88484f7a3dcf@google.com
Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-09-15 14:31:33 +00:00
|
|
|
|
2024-12-19 18:04:16 +01:00
|
|
|
|
2023-12-18 11:15:13 +05:30
|
|
|
|
rust: file: add abstraction for `poll_table`
The existing `CondVar` abstraction is a wrapper around
`wait_queue_head`, but it does not support all use-cases of the C
`wait_queue_head` type. To be specific, a `CondVar` cannot be registered
with a `struct poll_table`. This limitation has the advantage that you
do not need to call `synchronize_rcu` when destroying a `CondVar`.
However, we need the ability to register a `poll_table` with a
`wait_queue_head` in Rust Binder. To enable this, introduce a type
called `PollCondVar`, which is like `CondVar` except that you can
register a `poll_table`. We also introduce `PollTable`, which is a safe
wrapper around `poll_table` that is intended to be used with
`PollCondVar`.
The destructor of `PollCondVar` unconditionally calls `synchronize_rcu`
to ensure that the removal of epoll waiters has fully completed before
the `wait_queue_head` is destroyed.
That said, `synchronize_rcu` is rather expensive and is not needed in
all cases: If we have never registered a `poll_table` with the
`wait_queue_head`, then we don't need to call `synchronize_rcu`. (And
this is a common case in Binder - not all processes use Binder with
epoll.) The current implementation does not account for this, but if we
find that it is necessary to improve this, a future patch could store a
boolean next to the `wait_queue_head` to keep track of whether a
`poll_table` has ever been registered.
Reviewed-by: Benno Lossin <benno.lossin@proton.me>
Reviewed-by: Martin Rodriguez Reboredo <yakoyoku@gmail.com>
Reviewed-by: Trevor Gross <tmgross@umich.edu>
Signed-off-by: Alice Ryhl <aliceryhl@google.com>
Link: https://lore.kernel.org/r/20240915-alice-file-v10-8-88484f7a3dcf@google.com
Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-09-15 14:31:34 +00:00
|
|
|
|
2025-01-13 16:52:59 +05:30
|
|
|
|
2025-10-16 15:38:02 +02:00
|
|
|
|
rust: add find_bit_benchmark_rust module.
Microbenchmark protected by a config FIND_BIT_BENCHMARK_RUST,
following `find_bit_benchmark.c` but testing the Rust Bitmap API.
We add a fill_random() method protected by the config in order to
maintain the abstraction.
The sample output from the benchmark, both C and Rust version:
find_bit_benchmark.c output:
```
Start testing find_bit() with random-filled bitmap
[ 438.101937] find_next_bit: 860188 ns, 163419 iterations
[ 438.109471] find_next_zero_bit: 912342 ns, 164262 iterations
[ 438.116820] find_last_bit: 726003 ns, 163419 iterations
[ 438.130509] find_nth_bit: 7056993 ns, 16269 iterations
[ 438.139099] find_first_bit: 1963272 ns, 16270 iterations
[ 438.173043] find_first_and_bit: 27314224 ns, 32654 iterations
[ 438.180065] find_next_and_bit: 398752 ns, 73705 iterations
[ 438.186689]
Start testing find_bit() with sparse bitmap
[ 438.193375] find_next_bit: 9675 ns, 656 iterations
[ 438.201765] find_next_zero_bit: 1766136 ns, 327025 iterations
[ 438.208429] find_last_bit: 9017 ns, 656 iterations
[ 438.217816] find_nth_bit: 2749742 ns, 655 iterations
[ 438.225168] find_first_bit: 721799 ns, 656 iterations
[ 438.231797] find_first_and_bit: 2819 ns, 1 iterations
[ 438.238441] find_next_and_bit: 3159 ns, 1 iterations
```
find_bit_benchmark_rust.rs output:
```
[ 451.182459] find_bit_benchmark_rust:
[ 451.186688] Start testing find_bit() Rust with random-filled bitmap
[ 451.194450] next_bit: 777950 ns, 163644 iterations
[ 451.201997] next_zero_bit: 918889 ns, 164036 iterations
[ 451.208642] Start testing find_bit() Rust with sparse bitmap
[ 451.214300] next_bit: 9181 ns, 654 iterations
[ 451.222806] next_zero_bit: 1855504 ns, 327026 iterations
```
Here are the results from 32 samples, with 95% confidence interval.
The microbenchmark was built with RUST_BITMAP_HARDENED=n and run on a
machine that did not execute other processes.
Random-filled bitmap:
+-----------+-------+-----------+--------------+-----------+-----------+
| Benchmark | Lang | Mean (ms) | Std Dev (ms) | 95% CI Lo | 95% CI Hi |
+-----------+-------+-----------+--------------+-----------+-----------+
| find_bit/ | C | 825.07 | 53.89 | 806.40 | 843.74 |
| next_bit | Rust | 870.91 | 46.29 | 854.88 | 886.95 |
+-----------+-------+-----------+--------------+-----------+-----------+
| find_zero/| C | 933.56 | 56.34 | 914.04 | 953.08 |
| next_zero | Rust | 945.85 | 60.44 | 924.91 | 966.79 |
+-----------+-------+-----------+--------------+-----------+-----------+
Rust appears 5.5% slower for next_bit, 1.3% slower for next_zero.
Sparse bitmap:
+-----------+-------+-----------+--------------+-----------+-----------+
| Benchmark | Lang | Mean (ms) | Std Dev (ms) | 95% CI Lo | 95% CI Hi |
+-----------+-------+-----------+--------------+-----------+-----------+
| find_bit/ | C | 13.17 | 6.21 | 11.01 | 15.32 |
| next_bit | Rust | 14.30 | 8.27 | 11.43 | 17.17 |
+-----------+-------+-----------+--------------+-----------+-----------+
| find_zero/| C | 1859.31 | 82.30 | 1830.80 | 1887.83 |
| next_zero | Rust | 1908.09 | 139.82 | 1859.65 | 1956.54 |
+-----------+-------+-----------+--------------+-----------+-----------+
Rust appears 8.5% slower for next_bit, 2.6% slower for next_zero.
In summary, taking the arithmetic mean of all slow-downs, we can say
the Rust API has a 4.5% slowdown.
Suggested-by: Alice Ryhl <aliceryhl@google.com>
Suggested-by: Yury Norov (NVIDIA) <yury.norov@gmail.com>
Reviewed-by: Yury Norov (NVIDIA) <yury.norov@gmail.com>
Reviewed-by: Alice Ryhl <aliceryhl@google.com>
Signed-off-by: Burak Emir <bqe@google.com>
Signed-off-by: Yury Norov (NVIDIA) <yury.norov@gmail.com>
2025-09-08 07:21:54 +00:00
|
|
|
|
2022-12-28 06:03:40 +00:00
|
|
|
|
2025-07-14 15:52:04 -03:00
|
|
|
|
2023-04-11 02:45:39 -03:00
|
|
|
|
2024-09-15 14:31:31 +00:00
|
|
|
|
2024-02-16 17:27:23 +02:00
|
|
|
|
2025-12-26 20:17:07 +00:00
|
|
|
|
rust_binder: add Rust Binder driver
We're generally not proponents of rewrites (nasty uncomfortable things
that make you late for dinner!). So why rewrite Binder?
Binder has been evolving over the past 15+ years to meet the evolving
needs of Android. Its responsibilities, expectations, and complexity
have grown considerably during that time. While we expect Binder to
continue to evolve along with Android, there are a number of factors
that currently constrain our ability to develop/maintain it. Briefly
those are:
1. Complexity: Binder is at the intersection of everything in Android and
fulfills many responsibilities beyond IPC. It has become many things
to many people, and due to its many features and their interactions
with each other, its complexity is quite high. In just 6kLOC it must
deliver transactions to the right threads. It must correctly parse
and translate the contents of transactions, which can contain several
objects of different types (e.g., pointers, fds) that can interact
with each other. It controls the size of thread pools in userspace,
and ensures that transactions are assigned to threads in ways that
avoid deadlocks where the threadpool has run out of threads. It must
track refcounts of objects that are shared by several processes by
forwarding refcount changes between the processes correctly. It must
handle numerous error scenarios and it combines/nests 13 different
locks, 7 reference counters, and atomic variables. Finally, It must
do all of this as fast and efficiently as possible. Minor performance
regressions can cause a noticeably degraded user experience.
2. Things to improve: Thousand-line functions [1], error-prone error
handling [2], and confusing structure can occur as a code base grows
organically. After more than a decade of development, this codebase
could use an overhaul.
[1]: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/android/binder.c?h=v6.5#n2896
[2]: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/android/binder.c?h=v6.5#n3658
3. Security critical: Binder is a critical part of Android's sandboxing
strategy. Even Android's most de-privileged sandboxes (e.g. the
Chrome renderer, or SW Codec) have direct access to Binder. More than
just about any other component, it's important that Binder provide
robust security, and itself be robust against security
vulnerabilities.
It's #1 (high complexity) that has made continuing to evolve Binder and
resolving #2 (tech debt) exceptionally difficult without causing #3
(security issues). For Binder to continue to meet Android's needs, we
need better ways to manage (and reduce!) complexity without increasing
the risk.
The biggest change is obviously the choice of programming language. We
decided to use Rust because it directly addresses a number of the
challenges within Binder that we have faced during the last years. It
prevents mistakes with ref counting, locking, bounds checking, and also
does a lot to reduce the complexity of error handling. Additionally,
we've been able to use the more expressive type system to encode the
ownership semantics of the various structs and pointers, which takes the
complexity of managing object lifetimes out of the hands of the
programmer, reducing the risk of use-after-frees and similar problems.
Rust has many different pointer types that it uses to encode ownership
semantics into the type system, and this is probably one of the most
important aspects of how it helps in Binder. The Binder driver has a lot
of different objects that have complex ownership semantics; some
pointers own a refcount, some pointers have exclusive ownership, and
some pointers just reference the object and it is kept alive in some
other manner. With Rust, we can use a different pointer type for each
kind of pointer, which enables the compiler to enforce that the
ownership semantics are implemented correctly.
Another useful feature is Rust's error handling. Rust allows for more
simplified error handling with features such as destructors, and you get
compilation failures if errors are not properly handled. This means that
even though Rust requires you to spend more lines of code than C on
things such as writing down invariants that are left implicit in C, the
Rust driver is still slightly smaller than C binder: Rust is 5.5kLOC and
C is 5.8kLOC. (These numbers are excluding blank lines, comments,
binderfs, and any debugging facilities in C that are not yet implemented
in the Rust driver. The numbers include abstractions in rust/kernel/
that are unlikely to be used by other drivers than Binder.)
Although this rewrite completely rethinks how the code is structured and
how assumptions are enforced, we do not fundamentally change *how* the
driver does the things it does. A lot of careful thought has gone into
the existing design. The rewrite is aimed rather at improving code
health, structure, readability, robustness, security, maintainability
and extensibility. We also include more inline documentation, and
improve how assumptions in the code are enforced. Furthermore, all
unsafe code is annotated with a SAFETY comment that explains why it is
correct.
We have left the binderfs filesystem component in C. Rewriting it in
Rust would be a large amount of work and requires a lot of bindings to
the file system interfaces. Binderfs has not historically had the same
challenges with security and complexity, so rewriting binderfs seems to
have lower value than the rest of Binder.
Correctness and feature parity
------------------------------
Rust binder passes all tests that validate the correctness of Binder in
the Android Open Source Project. We can boot a device, and run a variety
of apps and functionality without issues. We have performed this both on
the Cuttlefish Android emulator device, and on a Pixel 6 Pro.
As for feature parity, Rust binder currently implements all features
that C binder supports, with the exception of some debugging facilities.
The missing debugging facilities will be added before we submit the Rust
implementation upstream.
Tracepoints
-----------
I did not include all of the tracepoints as I felt that the mechansim
for making C access fields of Rust structs should be discussed on list
separately. I also did not include the support for building Rust Binder
as a module since that requires exporting a bunch of additional symbols
on the C side.
Original RFC Link with old benchmark numbers:
https://lore.kernel.org/r/20231101-rust-binder-v1-0-08ba9197f637@google.com
Co-developed-by: Wedson Almeida Filho <wedsonaf@gmail.com>
Signed-off-by: Wedson Almeida Filho <wedsonaf@gmail.com>
Co-developed-by: Matt Gilbride <mattgilbride@google.com>
Signed-off-by: Matt Gilbride <mattgilbride@google.com>
Acked-by: Carlos Llamas <cmllamas@google.com>
Acked-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: Alice Ryhl <aliceryhl@google.com>
Link: https://lore.kernel.org/r/20250919-rust-binder-v2-1-a384b09f28dd@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2025-09-19 06:42:07 +00:00
|
|
|
|
2024-10-30 16:04:25 +00:00
|
|
|
|
2025-10-08 14:10:28 +02:00
|
|
|
|
2024-02-16 17:27:23 +02:00
|
|
|
|
2023-08-28 10:48:02 +00:00
|
|
|
|
2025-04-23 09:54:38 -04:00
|
|
|
|
2024-10-30 16:04:26 +00:00
|
|
|
|
2022-07-25 23:46:47 +02:00
|
|
|
|
rust: device: introduce Device::drvdata()
In C dev_get_drvdata() has specific requirements under which it is valid
to access the returned pointer. That is, drivers have to ensure that
(1) for the duration the returned pointer is accessed the driver is
bound and remains to be bound to the corresponding device,
(2) the returned void * is treated according to the driver's private
data type, i.e. according to what has been passed to
dev_set_drvdata().
In Rust, (1) can be ensured by simply requiring the Bound device
context, i.e. provide the drvdata() method for Device<Bound> only.
For (2) we would usually make the device type generic over the driver
type, e.g. Device<T: Driver>, where <T as Driver>::Data is the type of
the driver's private data.
However, a device does not have a driver type known at compile time and
may be bound to multiple drivers throughout its lifetime.
Hence, in order to be able to provide a safe accessor for the driver's
device private data, we have to do the type check on runtime.
This is achieved by letting a driver assert the expected type, which is
then compared to a type hash stored in struct device_private when
dev_set_drvdata() is called.
Example:
// `dev` is a `&Device<Bound>`.
let data = dev.drvdata::<SampleDriver>()?;
There are two aspects to note:
(1) Technically, the same check could be achieved by comparing the
struct device_driver pointer of struct device with the struct
device_driver pointer of the driver struct (e.g. struct
pci_driver).
However, this would - in addition the pointer comparison - require
to tie back the private driver data type to the struct
device_driver pointer of the driver struct to prove correctness.
Besides that, accessing the driver struct (stored in the module
structure) isn't trivial and would result into horrible code and
API ergonomics.
(2) Having a direct accessor to the driver's private data is not
commonly required (at least in Rust): Bus callback methods already
provide access to the driver's device private data through a &self
argument, while other driver entry points such as IRQs,
workqueues, timers, IOCTLs, etc. have their own private data with
separate ownership and lifetime.
In other words, a driver's device private data is only relevant
for driver model contexts (such a file private is only relevant
for file contexts).
Having that said, the motivation for accessing the driver's device
private data with Device<Bound>::drvdata() are interactions between
drivers. For instance, when an auxiliary driver calls back into its
parent, the parent has to be capable to derive its private data from the
corresponding device (i.e. the parent of the auxiliary device).
Reviewed-by: Alice Ryhl <aliceryhl@google.com>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ * Remove unnecessary `const _: ()` block,
* rename type_id_{store,match}() to {set,match}_type_id(),
* assert size_of::<bindings::driver_type>() >= size_of::<TypeId>(),
* add missing check in case Device::drvdata() is called from probe().
- Danilo ]
Signed-off-by: Danilo Krummrich <dakr@kernel.org>
2025-10-21 00:34:24 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2025-03-03 08:45:16 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2022-07-25 23:46:47 +02:00
|
|
|
|
2023-11-04 14:56:56 +00:00
|
|
|
|
2025-08-11 12:31:50 +00:00
|
|
|
|
2024-06-11 13:45:49 +02:00
|
|
|
|
2024-03-27 22:35:58 -03:00
|
|
|
|
2023-11-04 14:56:56 +00:00
|
|
|
|
2024-03-27 22:35:58 -03:00
|
|
|
|
|
|
|
|
|
2023-11-04 14:56:56 +00:00
|
|
|
|
2024-06-07 08:23:41 +00:00
|
|
|
|
2024-10-04 17:41:14 +02:00
|
|
|
|
2024-06-28 11:11:52 +02:00
|
|
|
|
2025-04-11 01:55:26 +02:00
|
|
|
|
2025-04-23 09:54:38 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
mm: declare VMA flags by bit
Patch series "initial work on making VMA flags a bitmap", v3.
We are in the rather silly situation that we are running out of VMA flags
as they are currently limited to a system word in size.
This leads to absurd situations where we limit features to 64-bit
architectures only because we simply do not have the ability to add a flag
for 32-bit ones.
This is very constraining and leads to hacks or, in the worst case, simply
an inability to implement features we want for entirely arbitrary reasons.
This also of course gives us something of a Y2K type situation in mm where
we might eventually exhaust all of the VMA flags even on 64-bit systems.
This series lays the groundwork for getting away from this limitation by
establishing VMA flags as a bitmap whose size we can increase in future
beyond 64 bits if required.
This is necessarily a highly iterative process given the extensive use of
VMA flags throughout the kernel, so we start by performing basic steps.
Firstly, we declare VMA flags by bit number rather than by value,
retaining the VM_xxx fields but in terms of these newly introduced
VMA_xxx_BIT fields.
While we are here, we use sparse annotations to ensure that, when dealing
with VMA bit number parameters, we cannot be passed values which are not
declared as such - providing some useful type safety.
We then introduce an opaque VMA flag type, much like the opaque mm_struct
flag type introduced in commit bb6525f2f8c4 ("mm: add bitmap mm->flags
field"), which we establish in union with vma->vm_flags (but still set at
system word size meaning there is no functional or data type size change).
We update the vm_flags_xxx() helpers to use this new bitmap, introducing
sensible helpers to do so.
This series lays the foundation for further work to expand the use of
bitmap VMA flags and eventually eliminate these arbitrary restrictions.
This patch (of 4):
In order to lay the groundwork for VMA flags being a bitmap rather than a
system word in size, we need to be able to consistently refer to VMA flags
by bit number rather than value.
Take this opportunity to do so in an enum which we which is additionally
useful for tooling to extract metadata from.
This additionally makes it very clear which bits are being used for what
at a glance.
We use the VMA_ prefix for the bit values as it is logical to do so since
these reference VMAs. We consistently suffix with _BIT to make it clear
what the values refer to.
We declare bit values even when the flags that use them would not be
enabled by config options as this is simply clearer and clearly defines
what bit numbers are used for what, at no additional cost.
We declare a sparse-bitwise type vma_flag_t which ensures that users can't
pass around invalid VMA flags by accident and prepares for future work
towards VMA flags being a bitmap where we want to ensure bit values are
type safe.
To make life easier, we declare some macro helpers - DECLARE_VMA_BIT()
allows us to avoid duplication in the enum bit number declarations (and
maintaining the sparse __bitwise attribute), and INIT_VM_FLAG() is used to
assist with declaration of flags.
Unfortunately we can't declare both in the enum, as we run into issue with
logic in the kernel requiring that flags are preprocessor definitions, and
additionally we cannot have a macro which declares another macro so we
must define each flag macro directly.
Additionally, update the VMA userland testing vma_internal.h header to
include these changes.
We also have to fix the parameters to the vma_flag_*_atomic() functions
since VMA_MAYBE_GUARD_BIT is now of type vma_flag_t and sparse will
complain otherwise.
We have to update some rather silly if-deffery found in mm/task_mmu.c
which would otherwise break.
Finally, we update the rust binding helper as now it cannot auto-detect
the flags at all.
Link: https://lkml.kernel.org/r/cover.1764064556.git.lorenzo.stoakes@oracle.com
Link: https://lkml.kernel.org/r/3a35e5a0bcfa00e84af24cbafc0653e74deda64a.1764064556.git.lorenzo.stoakes@oracle.com
Signed-off-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Reviewed-by: Pedro Falcato <pfalcato@suse.de>
Acked-by: Alice Ryhl <aliceryhl@google.com> [rust]
Cc: Alex Gaynor <alex.gaynor@gmail.com>
Cc: Alistair Popple <apopple@nvidia.com>
Cc: Andreas Hindborg <a.hindborg@kernel.org>
Cc: Axel Rasmussen <axelrasmussen@google.com>
Cc: Baolin Wang <baolin.wang@linux.alibaba.com>
Cc: Baoquan He <bhe@redhat.com>
Cc: Barry Song <baohua@kernel.org>
Cc: Ben Segall <bsegall@google.com>
Cc: Björn Roy Baron <bjorn3_gh@protonmail.com>
Cc: Boqun Feng <boqun.feng@gmail.com>
Cc: Byungchul Park <byungchul@sk.com>
Cc: Chengming Zhou <chengming.zhou@linux.dev>
Cc: Chris Li <chrisl@kernel.org>
Cc: Danilo Krummrich <dakr@kernel.org>
Cc: David Hildenbrand <david@redhat.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Dev Jain <dev.jain@arm.com>
Cc: Dietmar Eggemann <dietmar.eggemann@arm.com>
Cc: Gary Guo <gary@garyguo.net>
Cc: Gregory Price <gourry@gourry.net>
Cc: "Huang, Ying" <ying.huang@linux.alibaba.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jann Horn <jannh@google.com>
Cc: Jason Gunthorpe <jgg@ziepe.ca>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: Joshua Hahn <joshua.hahnjy@gmail.com>
Cc: Juri Lelli <juri.lelli@redhat.com>
Cc: Kairui Song <kasong@tencent.com>
Cc: Kees Cook <kees@kernel.org>
Cc: Kemeng Shi <shikemeng@huaweicloud.com>
Cc: Lance Yang <lance.yang@linux.dev>
Cc: Leon Romanovsky <leon@kernel.org>
Cc: Liam Howlett <liam.howlett@oracle.com>
Cc: Mathew Brost <matthew.brost@intel.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Mel Gorman <mgorman <mgorman@suse.de>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Miguel Ojeda <ojeda@kernel.org>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: Muchun Song <muchun.song@linux.dev>
Cc: Nhat Pham <nphamcs@gmail.com>
Cc: Nico Pache <npache@redhat.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Peter Xu <peterx@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Qi Zheng <zhengqi.arch@bytedance.com>
Cc: Rakie Kim <rakie.kim@sk.com>
Cc: Rik van Riel <riel@surriel.com>
Cc: Ryan Roberts <ryan.roberts@arm.com>
Cc: Shakeel Butt <shakeel.butt@linux.dev>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Suren Baghdasaryan <surenb@google.com>
Cc: Trevor Gross <tmgross@umich.edu>
Cc: Valentin Schneider <vschneid@redhat.com>
Cc: Vincent Guittot <vincent.guittot@linaro.org>
Cc: Wei Xu <weixugc@google.com>
Cc: xu xin <xu.xin16@zte.com.cn>
Cc: Yuanchu Xie <yuanchu@google.com>
Cc: Zi Yan <ziy@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
2025-11-25 10:00:59 +00:00
|
|
|
|
mm/ksm: fix flag-dropping behavior in ksm_madvise
syzkaller discovered the following crash: (kernel BUG)
[ 44.607039] ------------[ cut here ]------------
[ 44.607422] kernel BUG at mm/userfaultfd.c:2067!
[ 44.608148] Oops: invalid opcode: 0000 [#1] SMP DEBUG_PAGEALLOC KASAN NOPTI
[ 44.608814] CPU: 1 UID: 0 PID: 2475 Comm: reproducer Not tainted 6.16.0-rc6 #1 PREEMPT(none)
[ 44.609635] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.16.3-0-ga6ed6b701f0a-prebuilt.qemu.org 04/01/2014
[ 44.610695] RIP: 0010:userfaultfd_release_all+0x3a8/0x460
<snip other registers, drop unreliable trace>
[ 44.617726] Call Trace:
[ 44.617926] <TASK>
[ 44.619284] userfaultfd_release+0xef/0x1b0
[ 44.620976] __fput+0x3f9/0xb60
[ 44.621240] fput_close_sync+0x110/0x210
[ 44.622222] __x64_sys_close+0x8f/0x120
[ 44.622530] do_syscall_64+0x5b/0x2f0
[ 44.622840] entry_SYSCALL_64_after_hwframe+0x76/0x7e
[ 44.623244] RIP: 0033:0x7f365bb3f227
Kernel panics because it detects UFFD inconsistency during
userfaultfd_release_all(). Specifically, a VMA which has a valid pointer
to vma->vm_userfaultfd_ctx, but no UFFD flags in vma->vm_flags.
The inconsistency is caused in ksm_madvise(): when user calls madvise()
with MADV_UNMEARGEABLE on a VMA that is registered for UFFD in MINOR mode,
it accidentally clears all flags stored in the upper 32 bits of
vma->vm_flags.
Assuming x86_64 kernel build, unsigned long is 64-bit and unsigned int and
int are 32-bit wide. This setup causes the following mishap during the &=
~VM_MERGEABLE assignment.
VM_MERGEABLE is a 32-bit constant of type unsigned int, 0x8000'0000.
After ~ is applied, it becomes 0x7fff'ffff unsigned int, which is then
promoted to unsigned long before the & operation. This promotion fills
upper 32 bits with leading 0s, as we're doing unsigned conversion (and
even for a signed conversion, this wouldn't help as the leading bit is 0).
& operation thus ends up AND-ing vm_flags with 0x0000'0000'7fff'ffff
instead of intended 0xffff'ffff'7fff'ffff and hence accidentally clears
the upper 32-bits of its value.
Fix it by changing `VM_MERGEABLE` constant to unsigned long, using the
BIT() macro.
Note: other VM_* flags are not affected: This only happens to the
VM_MERGEABLE flag, as the other VM_* flags are all constants of type int
and after ~ operation, they end up with leading 1 and are thus converted
to unsigned long with leading 1s.
Note 2:
After commit 31defc3b01d9 ("userfaultfd: remove (VM_)BUG_ON()s"), this is
no longer a kernel BUG, but a WARNING at the same place:
[ 45.595973] WARNING: CPU: 1 PID: 2474 at mm/userfaultfd.c:2067
but the root-cause (flag-drop) remains the same.
[akpm@linux-foundation.org: rust bindgen wasn't able to handle BIT(), from Miguel]
Link: https://lore.kernel.org/oe-kbuild-all/202510030449.VfSaAjvd-lkp@intel.com/
Link: https://lkml.kernel.org/r/20251001090353.57523-2-acsjakub@amazon.de
Fixes: 7677f7fd8be7 ("userfaultfd: add minor fault registration mode")
Signed-off-by: Jakub Acs <acsjakub@amazon.de>
Signed-off-by: Miguel Ojeda <miguel.ojeda.sandonis@gmail.com>
Acked-by: David Hildenbrand <david@redhat.com>
Acked-by: SeongJae Park <sj@kernel.org>
Tested-by: Alice Ryhl <aliceryhl@google.com>
Tested-by: Miguel Ojeda <miguel.ojeda.sandonis@gmail.com>
Cc: Xu Xin <xu.xin16@zte.com.cn>
Cc: Chengming Zhou <chengming.zhou@linux.dev>
Cc: Peter Xu <peterx@redhat.com>
Cc: Axel Rasmussen <axelrasmussen@google.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
2025-10-01 09:03:52 +00:00
|
|
|
|
mm: declare VMA flags by bit
Patch series "initial work on making VMA flags a bitmap", v3.
We are in the rather silly situation that we are running out of VMA flags
as they are currently limited to a system word in size.
This leads to absurd situations where we limit features to 64-bit
architectures only because we simply do not have the ability to add a flag
for 32-bit ones.
This is very constraining and leads to hacks or, in the worst case, simply
an inability to implement features we want for entirely arbitrary reasons.
This also of course gives us something of a Y2K type situation in mm where
we might eventually exhaust all of the VMA flags even on 64-bit systems.
This series lays the groundwork for getting away from this limitation by
establishing VMA flags as a bitmap whose size we can increase in future
beyond 64 bits if required.
This is necessarily a highly iterative process given the extensive use of
VMA flags throughout the kernel, so we start by performing basic steps.
Firstly, we declare VMA flags by bit number rather than by value,
retaining the VM_xxx fields but in terms of these newly introduced
VMA_xxx_BIT fields.
While we are here, we use sparse annotations to ensure that, when dealing
with VMA bit number parameters, we cannot be passed values which are not
declared as such - providing some useful type safety.
We then introduce an opaque VMA flag type, much like the opaque mm_struct
flag type introduced in commit bb6525f2f8c4 ("mm: add bitmap mm->flags
field"), which we establish in union with vma->vm_flags (but still set at
system word size meaning there is no functional or data type size change).
We update the vm_flags_xxx() helpers to use this new bitmap, introducing
sensible helpers to do so.
This series lays the foundation for further work to expand the use of
bitmap VMA flags and eventually eliminate these arbitrary restrictions.
This patch (of 4):
In order to lay the groundwork for VMA flags being a bitmap rather than a
system word in size, we need to be able to consistently refer to VMA flags
by bit number rather than value.
Take this opportunity to do so in an enum which we which is additionally
useful for tooling to extract metadata from.
This additionally makes it very clear which bits are being used for what
at a glance.
We use the VMA_ prefix for the bit values as it is logical to do so since
these reference VMAs. We consistently suffix with _BIT to make it clear
what the values refer to.
We declare bit values even when the flags that use them would not be
enabled by config options as this is simply clearer and clearly defines
what bit numbers are used for what, at no additional cost.
We declare a sparse-bitwise type vma_flag_t which ensures that users can't
pass around invalid VMA flags by accident and prepares for future work
towards VMA flags being a bitmap where we want to ensure bit values are
type safe.
To make life easier, we declare some macro helpers - DECLARE_VMA_BIT()
allows us to avoid duplication in the enum bit number declarations (and
maintaining the sparse __bitwise attribute), and INIT_VM_FLAG() is used to
assist with declaration of flags.
Unfortunately we can't declare both in the enum, as we run into issue with
logic in the kernel requiring that flags are preprocessor definitions, and
additionally we cannot have a macro which declares another macro so we
must define each flag macro directly.
Additionally, update the VMA userland testing vma_internal.h header to
include these changes.
We also have to fix the parameters to the vma_flag_*_atomic() functions
since VMA_MAYBE_GUARD_BIT is now of type vma_flag_t and sparse will
complain otherwise.
We have to update some rather silly if-deffery found in mm/task_mmu.c
which would otherwise break.
Finally, we update the rust binding helper as now it cannot auto-detect
the flags at all.
Link: https://lkml.kernel.org/r/cover.1764064556.git.lorenzo.stoakes@oracle.com
Link: https://lkml.kernel.org/r/3a35e5a0bcfa00e84af24cbafc0653e74deda64a.1764064556.git.lorenzo.stoakes@oracle.com
Signed-off-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Reviewed-by: Pedro Falcato <pfalcato@suse.de>
Acked-by: Alice Ryhl <aliceryhl@google.com> [rust]
Cc: Alex Gaynor <alex.gaynor@gmail.com>
Cc: Alistair Popple <apopple@nvidia.com>
Cc: Andreas Hindborg <a.hindborg@kernel.org>
Cc: Axel Rasmussen <axelrasmussen@google.com>
Cc: Baolin Wang <baolin.wang@linux.alibaba.com>
Cc: Baoquan He <bhe@redhat.com>
Cc: Barry Song <baohua@kernel.org>
Cc: Ben Segall <bsegall@google.com>
Cc: Björn Roy Baron <bjorn3_gh@protonmail.com>
Cc: Boqun Feng <boqun.feng@gmail.com>
Cc: Byungchul Park <byungchul@sk.com>
Cc: Chengming Zhou <chengming.zhou@linux.dev>
Cc: Chris Li <chrisl@kernel.org>
Cc: Danilo Krummrich <dakr@kernel.org>
Cc: David Hildenbrand <david@redhat.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Dev Jain <dev.jain@arm.com>
Cc: Dietmar Eggemann <dietmar.eggemann@arm.com>
Cc: Gary Guo <gary@garyguo.net>
Cc: Gregory Price <gourry@gourry.net>
Cc: "Huang, Ying" <ying.huang@linux.alibaba.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jann Horn <jannh@google.com>
Cc: Jason Gunthorpe <jgg@ziepe.ca>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: Joshua Hahn <joshua.hahnjy@gmail.com>
Cc: Juri Lelli <juri.lelli@redhat.com>
Cc: Kairui Song <kasong@tencent.com>
Cc: Kees Cook <kees@kernel.org>
Cc: Kemeng Shi <shikemeng@huaweicloud.com>
Cc: Lance Yang <lance.yang@linux.dev>
Cc: Leon Romanovsky <leon@kernel.org>
Cc: Liam Howlett <liam.howlett@oracle.com>
Cc: Mathew Brost <matthew.brost@intel.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Mel Gorman <mgorman <mgorman@suse.de>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Miguel Ojeda <ojeda@kernel.org>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: Muchun Song <muchun.song@linux.dev>
Cc: Nhat Pham <nphamcs@gmail.com>
Cc: Nico Pache <npache@redhat.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Peter Xu <peterx@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Qi Zheng <zhengqi.arch@bytedance.com>
Cc: Rakie Kim <rakie.kim@sk.com>
Cc: Rik van Riel <riel@surriel.com>
Cc: Ryan Roberts <ryan.roberts@arm.com>
Cc: Shakeel Butt <shakeel.butt@linux.dev>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Suren Baghdasaryan <surenb@google.com>
Cc: Trevor Gross <tmgross@umich.edu>
Cc: Valentin Schneider <vschneid@redhat.com>
Cc: Vincent Guittot <vincent.guittot@linaro.org>
Cc: Wei Xu <weixugc@google.com>
Cc: xu xin <xu.xin16@zte.com.cn>
Cc: Yuanchu Xie <yuanchu@google.com>
Cc: Zi Yan <ziy@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
2025-11-25 10:00:59 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
rust_binder: add Rust Binder driver
We're generally not proponents of rewrites (nasty uncomfortable things
that make you late for dinner!). So why rewrite Binder?
Binder has been evolving over the past 15+ years to meet the evolving
needs of Android. Its responsibilities, expectations, and complexity
have grown considerably during that time. While we expect Binder to
continue to evolve along with Android, there are a number of factors
that currently constrain our ability to develop/maintain it. Briefly
those are:
1. Complexity: Binder is at the intersection of everything in Android and
fulfills many responsibilities beyond IPC. It has become many things
to many people, and due to its many features and their interactions
with each other, its complexity is quite high. In just 6kLOC it must
deliver transactions to the right threads. It must correctly parse
and translate the contents of transactions, which can contain several
objects of different types (e.g., pointers, fds) that can interact
with each other. It controls the size of thread pools in userspace,
and ensures that transactions are assigned to threads in ways that
avoid deadlocks where the threadpool has run out of threads. It must
track refcounts of objects that are shared by several processes by
forwarding refcount changes between the processes correctly. It must
handle numerous error scenarios and it combines/nests 13 different
locks, 7 reference counters, and atomic variables. Finally, It must
do all of this as fast and efficiently as possible. Minor performance
regressions can cause a noticeably degraded user experience.
2. Things to improve: Thousand-line functions [1], error-prone error
handling [2], and confusing structure can occur as a code base grows
organically. After more than a decade of development, this codebase
could use an overhaul.
[1]: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/android/binder.c?h=v6.5#n2896
[2]: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/android/binder.c?h=v6.5#n3658
3. Security critical: Binder is a critical part of Android's sandboxing
strategy. Even Android's most de-privileged sandboxes (e.g. the
Chrome renderer, or SW Codec) have direct access to Binder. More than
just about any other component, it's important that Binder provide
robust security, and itself be robust against security
vulnerabilities.
It's #1 (high complexity) that has made continuing to evolve Binder and
resolving #2 (tech debt) exceptionally difficult without causing #3
(security issues). For Binder to continue to meet Android's needs, we
need better ways to manage (and reduce!) complexity without increasing
the risk.
The biggest change is obviously the choice of programming language. We
decided to use Rust because it directly addresses a number of the
challenges within Binder that we have faced during the last years. It
prevents mistakes with ref counting, locking, bounds checking, and also
does a lot to reduce the complexity of error handling. Additionally,
we've been able to use the more expressive type system to encode the
ownership semantics of the various structs and pointers, which takes the
complexity of managing object lifetimes out of the hands of the
programmer, reducing the risk of use-after-frees and similar problems.
Rust has many different pointer types that it uses to encode ownership
semantics into the type system, and this is probably one of the most
important aspects of how it helps in Binder. The Binder driver has a lot
of different objects that have complex ownership semantics; some
pointers own a refcount, some pointers have exclusive ownership, and
some pointers just reference the object and it is kept alive in some
other manner. With Rust, we can use a different pointer type for each
kind of pointer, which enables the compiler to enforce that the
ownership semantics are implemented correctly.
Another useful feature is Rust's error handling. Rust allows for more
simplified error handling with features such as destructors, and you get
compilation failures if errors are not properly handled. This means that
even though Rust requires you to spend more lines of code than C on
things such as writing down invariants that are left implicit in C, the
Rust driver is still slightly smaller than C binder: Rust is 5.5kLOC and
C is 5.8kLOC. (These numbers are excluding blank lines, comments,
binderfs, and any debugging facilities in C that are not yet implemented
in the Rust driver. The numbers include abstractions in rust/kernel/
that are unlikely to be used by other drivers than Binder.)
Although this rewrite completely rethinks how the code is structured and
how assumptions are enforced, we do not fundamentally change *how* the
driver does the things it does. A lot of careful thought has gone into
the existing design. The rewrite is aimed rather at improving code
health, structure, readability, robustness, security, maintainability
and extensibility. We also include more inline documentation, and
improve how assumptions in the code are enforced. Furthermore, all
unsafe code is annotated with a SAFETY comment that explains why it is
correct.
We have left the binderfs filesystem component in C. Rewriting it in
Rust would be a large amount of work and requires a lot of bindings to
the file system interfaces. Binderfs has not historically had the same
challenges with security and complexity, so rewriting binderfs seems to
have lower value than the rest of Binder.
Correctness and feature parity
------------------------------
Rust binder passes all tests that validate the correctness of Binder in
the Android Open Source Project. We can boot a device, and run a variety
of apps and functionality without issues. We have performed this both on
the Cuttlefish Android emulator device, and on a Pixel 6 Pro.
As for feature parity, Rust binder currently implements all features
that C binder supports, with the exception of some debugging facilities.
The missing debugging facilities will be added before we submit the Rust
implementation upstream.
Tracepoints
-----------
I did not include all of the tracepoints as I felt that the mechansim
for making C access fields of Rust structs should be discussed on list
separately. I also did not include the support for building Rust Binder
as a module since that requires exporting a bunch of additional symbols
on the C side.
Original RFC Link with old benchmark numbers:
https://lore.kernel.org/r/20231101-rust-binder-v1-0-08ba9197f637@google.com
Co-developed-by: Wedson Almeida Filho <wedsonaf@gmail.com>
Signed-off-by: Wedson Almeida Filho <wedsonaf@gmail.com>
Co-developed-by: Matt Gilbride <mattgilbride@google.com>
Signed-off-by: Matt Gilbride <mattgilbride@google.com>
Acked-by: Carlos Llamas <cmllamas@google.com>
Acked-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: Alice Ryhl <aliceryhl@google.com>
Link: https://lore.kernel.org/r/20250919-rust-binder-v2-1-a384b09f28dd@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2025-09-19 06:42:07 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|