linux-kernel - Re: [PATCH v2 3/8] rust: uaccess: add UserSliceWriter::write_slice

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <aPeWOhycOIl_rlI-@google.com>
Date: Tue, 21 Oct 2025 14:18:34 +0000
From: Alice Ryhl <aliceryhl@...gle.com>
To: Danilo Krummrich <dakr@...nel.org>
Cc: gregkh@...uxfoundation.org, rafael@...nel.org, ojeda@...nel.org, 
	alex.gaynor@...il.com, boqun.feng@...il.com, gary@...yguo.net, 
	bjorn3_gh@...tonmail.com, lossin@...nel.org, a.hindborg@...nel.org, 
	tmgross@...ch.edu, mmaurer@...gle.com, rust-for-linux@...r.kernel.org, 
	linux-fsdevel@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 3/8] rust: uaccess: add UserSliceWriter::write_slice_partial()

On Tue, Oct 21, 2025 at 04:14:22PM +0200, Danilo Krummrich wrote:
> On Tue Oct 21, 2025 at 4:00 PM CEST, Alice Ryhl wrote:
> > On Tue, Oct 21, 2025 at 12:26:15AM +0200, Danilo Krummrich wrote:
> >> The existing write_slice() method is a wrapper around copy_to_user() and
> >> expects the user buffer to be larger than the source buffer.
> >> 
> >> However, userspace may split up reads in multiple partial operations
> >> providing an offset into the source buffer and a smaller user buffer.
> >> 
> >> In order to support this common case, provide a helper for partial
> >> writes.
> >> 
> >> Signed-off-by: Danilo Krummrich <dakr@...nel.org>
> >>  rust/kernel/uaccess.rs | 24 ++++++++++++++++++++++++
> >>  1 file changed, 24 insertions(+)
> >> 
> >> diff --git a/rust/kernel/uaccess.rs b/rust/kernel/uaccess.rs
> >> index 2061a7e10c65..40d47e94b54f 100644
> >> --- a/rust/kernel/uaccess.rs
> >> +++ b/rust/kernel/uaccess.rs
> >> @@ -463,6 +463,30 @@ pub fn write_slice(&mut self, data: &[u8]) -> Result {
> >>          Ok(())
> >>      }
> >>  
> >> +    /// Writes raw data to this user pointer from a kernel buffer partially.
> >> +    ///
> >> +    /// This is the same as [`Self::write_slice`] but considers the given `offset` into `data` and
> >> +    /// truncates the write to the boundaries of `self` and `data`.
> >> +    ///
> >> +    /// On success, returns the number of bytes written.
> >> +    pub fn write_slice_partial(&mut self, data: &[u8], offset: file::Offset) -> Result<usize> {
> >
> > I think for the current function signature, it's kind of weird to take a
> > file::Offset parameter
> >
> > On one hand, it is described like a generic function for writing a
> > partial slice, and if that's what it is, then I would argue it should
> > take usize because it's an offset into the slice.
> >
> > On another hand, I think what you're actually trying to do is implement
> > the simple_[read_from|write_to]_buffer utilities for user slices, but
> > it's only a "partial" version of those utilities. The full utility takes
> > a `&mut loff_t` so that it can also perform the required modification to
> > the offset.
> 
> Originally, it was intended to be the latter. And, in fact, earlier code (that
> did not git the mailing list) had a &mut file::Offset argument (was &mut i64
> back then).
> 
> However, for the version I sent to the list I chose the former because I
> considered it to be more flexible.
> 
> Now, in v2, it's indeed a bit mixed up. I think what we should do is to have
> both
> 
> 	fn write_slice_partial(&mut self, data: &[u8], offset: usize) -> Result<usize>
> 
> and
> 
> 	fn write_slice_???(&mut self, data: &[u8], offset: &mut file::Offset) -> Result<usize>
> 
> which can forward to write_slice_partial() and update the buffer.

SGTM.

> Any name suggestions?

I would suggest keeping the name of the equivalent C method:
simple_read_from_buffer/simple_write_to_buffer

Alice