[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <Zw63vD879qeHUA_b@boqun-archlinux>
Date: Tue, 15 Oct 2024 11:43:08 -0700
From: Boqun Feng <boqun.feng@...il.com>
To: Alice Ryhl <alice@...l.io>
Cc: Alice Ryhl <aliceryhl@...gle.com>, Miguel Ojeda <ojeda@...nel.org>,
Andrew Morton <akpm@...ux-foundation.org>,
"Matthew Wilcox (Oracle)" <willy@...radead.org>,
Alex Gaynor <alex.gaynor@...il.com>, Gary Guo <gary@...yguo.net>,
Björn Roy Baron <bjorn3_gh@...tonmail.com>,
Benno Lossin <benno.lossin@...ton.me>,
Andreas Hindborg <a.hindborg@...nel.org>,
Trevor Gross <tmgross@...ch.edu>, linux-mm@...ck.org,
rust-for-linux@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] rust: page: add Rust version of PAGE_ALIGN
On Tue, Oct 15, 2024 at 08:44:22PM +0200, Alice Ryhl wrote:
> On 10/15/24 8:33 PM, Boqun Feng wrote:
> > On Tue, Oct 15, 2024 at 02:28:28PM +0000, Alice Ryhl wrote:
> > > This is a useful for helper for working with indices into buffers that
> > > consist of several pages. I forgot to include it when I added PAGE_SIZE
> > > and PAGE_MASK for the same purpose in commit fc6e66f4696b ("rust: add
> > > abstraction for `struct page`").
> > >
> > > Signed-off-by: Alice Ryhl <aliceryhl@...gle.com>
> > > ---
> > > rust/kernel/page.rs | 5 +++++
> > > 1 file changed, 5 insertions(+)
> > >
> > > diff --git a/rust/kernel/page.rs b/rust/kernel/page.rs
> > > index 208a006d587c..90846e3fe829 100644
> > > --- a/rust/kernel/page.rs
> > > +++ b/rust/kernel/page.rs
> > > @@ -20,6 +20,11 @@
> > > /// A bitmask that gives the page containing a given address.
> > > pub const PAGE_MASK: usize = !(PAGE_SIZE - 1);
> > > +/// Round up the given number to a multiple of `PAGE_SIZE`.
> > > +pub fn page_align(addr: usize) -> usize {
> >
> > Make it a `const` function? For example, there is a VMBUS_RING_SIZE()
> > macro in C that calculates a const number for a vmbus ringbuffer size,
> > so it will be useful.
>
> Good idea.
>
> > > + (addr + (PAGE_SIZE - 1)) & PAGE_MASK
> >
> > I guess overflows are unexpected, i.e. the users should not pass a
> > `addr` that `> usize::MAX - PAGE_SIZE + 1`?
> Correct. If this wraps around to zero, that's incorrect. Note that the minus
> one is in brackets to only trigger overflow detection in the right cases.
>
Make senses, with the const change and some docs on the overflow case,
feel free to add:
Reviewed-by: Boqun Feng <boqun.feng@...il.com>
Regards,
Boqun
> I can mention this in the docs.
>
> Alice
Powered by blists - more mailing lists