This constructor is useful when you just want to create a BitmapVec
without allocating but don't care how large it is.
Signed-off-by: Alice Ryhl <aliceryhl@google.com>
---
rust/kernel/bitmap.rs | 10 ++++++++++
1 file changed, 10 insertions(+)
diff --git a/rust/kernel/bitmap.rs b/rust/kernel/bitmap.rs
index 15fa23b45054b9272415fcc000e3e3b52c74d7c1..4ffe9eb0f208a3d62016e00297f5a0800aa33336 100644
--- a/rust/kernel/bitmap.rs
+++ b/rust/kernel/bitmap.rs
@@ -232,6 +232,16 @@ impl BitmapVec {
/// The maximum length that avoids allocating.
pub const NO_ALLOC_MAX_LEN: usize = BITS_PER_LONG;
+ /// Constructs a new [`BitmapVec`] without allocating.
+ #[inline]
+ pub fn new_small() -> Self {
+ // INVARIANT: `nbits <= NO_ALLOC_MAX_LEN`, so an inline bitmap is the right repr.
+ BitmapVec {
+ repr: BitmapRepr { bitmap: 0 },
+ nbits: BitmapVec::NO_ALLOC_MAX_LEN,
+ }
+ }
+
/// Constructs a new [`BitmapVec`].
///
/// Fails with [`AllocError`] when the [`BitmapVec`] could not be allocated. This
--
2.51.0.869.ge66316f041-goog
On Tue, Oct 21, 2025 at 01:32:44PM +0000, Alice Ryhl wrote:
> This constructor is useful when you just want to create a BitmapVec
> without allocating but don't care how large it is.
>
> Signed-off-by: Alice Ryhl <aliceryhl@google.com>
> ---
> rust/kernel/bitmap.rs | 10 ++++++++++
> 1 file changed, 10 insertions(+)
>
> diff --git a/rust/kernel/bitmap.rs b/rust/kernel/bitmap.rs
> index 15fa23b45054b9272415fcc000e3e3b52c74d7c1..4ffe9eb0f208a3d62016e00297f5a0800aa33336 100644
> --- a/rust/kernel/bitmap.rs
> +++ b/rust/kernel/bitmap.rs
> @@ -232,6 +232,16 @@ impl BitmapVec {
> /// The maximum length that avoids allocating.
> pub const NO_ALLOC_MAX_LEN: usize = BITS_PER_LONG;
>
> + /// Constructs a new [`BitmapVec`] without allocating.
> + #[inline]
> + pub fn new_small() -> Self {
Nit: maybe:
/// Construct a longest possible inline [`BitmapVec`].
#[inline]
pub fn new_inline() ...
This 'small vs large' lingo is internal to bitmaps. I don't think it
is worth to expose it in the interfaces. 'Inline' or 'inplace' sounds
better to me.
With that,
Acked-by: Yury Norov (NVIDIA) <yury.norov@gmail.com>
> + // INVARIANT: `nbits <= NO_ALLOC_MAX_LEN`, so an inline bitmap is the right repr.
> + BitmapVec {
> + repr: BitmapRepr { bitmap: 0 },
> + nbits: BitmapVec::NO_ALLOC_MAX_LEN,
A side note: after merging bitfields, we may switch inline bitmaps to to
bitfield!() {
0:31 nbits;
32:64 bitmap;
}
Thanks,
Yury
> + }
> + }
> +
> /// Constructs a new [`BitmapVec`].
> ///
> /// Fails with [`AllocError`] when the [`BitmapVec`] could not be allocated. This
>
> --
> 2.51.0.869.ge66316f041-goog
On Thu, Oct 23, 2025 at 01:33:03PM -0400, Yury Norov wrote:
> On Tue, Oct 21, 2025 at 01:32:44PM +0000, Alice Ryhl wrote:
> > This constructor is useful when you just want to create a BitmapVec
> > without allocating but don't care how large it is.
> >
> > Signed-off-by: Alice Ryhl <aliceryhl@google.com>
> > ---
> > rust/kernel/bitmap.rs | 10 ++++++++++
> > 1 file changed, 10 insertions(+)
> >
> > diff --git a/rust/kernel/bitmap.rs b/rust/kernel/bitmap.rs
> > index 15fa23b45054b9272415fcc000e3e3b52c74d7c1..4ffe9eb0f208a3d62016e00297f5a0800aa33336 100644
> > --- a/rust/kernel/bitmap.rs
> > +++ b/rust/kernel/bitmap.rs
> > @@ -232,6 +232,16 @@ impl BitmapVec {
> > /// The maximum length that avoids allocating.
> > pub const NO_ALLOC_MAX_LEN: usize = BITS_PER_LONG;
> >
> > + /// Constructs a new [`BitmapVec`] without allocating.
> > + #[inline]
> > + pub fn new_small() -> Self {
>
> Nit: maybe:
>
> /// Construct a longest possible inline [`BitmapVec`].
> #[inline]
> pub fn new_inline() ...
>
> This 'small vs large' lingo is internal to bitmaps. I don't think it
> is worth to expose it in the interfaces. 'Inline' or 'inplace' sounds
> better to me.
>
> With that,
>
> Acked-by: Yury Norov (NVIDIA) <yury.norov@gmail.com>
Makes sense. Will reword to 'inline', thanks!
> > + // INVARIANT: `nbits <= NO_ALLOC_MAX_LEN`, so an inline bitmap is the right repr.
> > + BitmapVec {
> > + repr: BitmapRepr { bitmap: 0 },
> > + nbits: BitmapVec::NO_ALLOC_MAX_LEN,
>
> A side note: after merging bitfields, we may switch inline bitmaps to to
>
> bitfield!() {
> 0:31 nbits;
> 32:64 bitmap;
> }
Personally I think I would prefer to keep the union only on the `repr`
field like it is now.
Alice
On Tue, Oct 21, 2025 at 3:33 PM Alice Ryhl <aliceryhl@google.com> wrote: > > This constructor is useful when you just want to create a BitmapVec > without allocating but don't care how large it is. > > Signed-off-by: Alice Ryhl <aliceryhl@google.com> Reviewed-by: Burak Emir <bqe@google.com>
© 2016 - 2026 Red Hat, Inc.