Adds just enough support to allow device drivers to feed entropy to the
central pool.

Signed-off-by: Matthew Maurer <mmaurer@google.com>
---
Changes in v3:
- Fixed doclink to be srctree-based instead of relative.
- Switched to prelude import instead of fine-grained imports.
- Link to v2: https://lore.kernel.org/r/20251216-add-entropy-v2-1-4d866f251474@google.com

Changes in v2:
- Added more details in the docs about the API, specifically about it
  not crediting entropy and when it ought to be used.
- Link to v1: https://lore.kernel.org/r/20251212-add-entropy-v1-1-e70ad1bc9c65@google.com
---
 rust/kernel/lib.rs  |  1 +
 rust/kernel/rand.rs | 18 ++++++++++++++++++
 2 files changed, 19 insertions(+)

diff --git a/rust/kernel/lib.rs b/rust/kernel/lib.rs
index f812cf12004286962985a068665443dc22c389a2..bf64752d276b0bdea06ac0de8a5e219190129377 100644
--- a/rust/kernel/lib.rs
+++ b/rust/kernel/lib.rs
@@ -130,6 +130,7 @@
 pub mod ptr;
 #[cfg(CONFIG_RUST_PWM_ABSTRACTIONS)]
 pub mod pwm;
+pub mod rand;
 pub mod rbtree;
 pub mod regulator;
 pub mod revocable;
diff --git a/rust/kernel/rand.rs b/rust/kernel/rand.rs
new file mode 100644
index 0000000000000000000000000000000000000000..3e0aef2f53af122f510b34c4743a20452aaad9ad
--- /dev/null
+++ b/rust/kernel/rand.rs
@@ -0,0 +1,18 @@
+// SPDX-License-Identifier: GPL-2.0
+
+//! Randomness.
+//!
+//! C header: [`include/linux/random.h`](srctree/include/linux/random.h)
+
+use crate::prelude::*;
+
+/// Adds the given buffer to the entropy pool, but does not credit any entropy.
+///
+/// This is intended for use mixing in data that is likely to differ between devices or boots, but
+/// may otherwise be predictable. Examples include MAC addresses or RTC values. This slightly
+/// improves randomness in entropy-constrained environments (especially common for embedded
+/// devices).
+pub fn add_device_randomness(buf: &[u8]) {
+    // SAFETY: We just need the pointer to be valid for the length, which a slice provides.
+    unsafe { bindings::add_device_randomness(buf.as_ptr().cast::<c_void>(), buf.len()) };
+}

---
base-commit: 008d3547aae5bc86fac3eda317489169c3fda112
change-id: 20251029-add-entropy-f57e12ebe110

Best regards,
-- 
Matthew Maurer <mmaurer@google.com>

On Fri, 26 Dec 2025 at 21:15, Matthew Maurer <mmaurer@google.com> wrote:
> diff --git a/rust/kernel/rand.rs b/rust/kernel/rand.rs
> new file mode 100644
> index 0000000000000000000000000000000000000000..3e0aef2f53af122f510b34c4743a20452aaad9ad
> --- /dev/null
> +++ b/rust/kernel/rand.rs
> @@ -0,0 +1,18 @@
> +// SPDX-License-Identifier: GPL-2.0
> +
> +//! Randomness.
> +//!
> +//! C header: [`include/linux/random.h`](srctree/include/linux/random.h)

So why rand and not random? If there is good reason to not use random.rs then
can you add that reason to commit message.

    Argillander

On Tue, Dec 30, 2025 at 5:54 AM Kari Argillander
<kari.argillander@gmail.com> wrote:
>
> On Fri, 26 Dec 2025 at 21:15, Matthew Maurer <mmaurer@google.com> wrote:
> > diff --git a/rust/kernel/rand.rs b/rust/kernel/rand.rs
> > new file mode 100644
> > index 0000000000000000000000000000000000000000..3e0aef2f53af122f510b34c4743a20452aaad9ad
> > --- /dev/null
> > +++ b/rust/kernel/rand.rs
> > @@ -0,0 +1,18 @@
> > +// SPDX-License-Identifier: GPL-2.0
> > +
> > +//! Randomness.
> > +//!
> > +//! C header: [`include/linux/random.h`](srctree/include/linux/random.h)
>
> So why rand and not random? If there is good reason to not use random.rs then
> can you add that reason to commit message.

I have no reason for rand over random and no strong preference. I'll
switch it to `random` in the next rev unless someone else chimes in.

>
>     Argillander

On Fri, Dec 26, 2025 at 07:15:10PM +0000, Matthew Maurer wrote:
> Adds just enough support to allow device drivers to feed entropy to the
> central pool.
> 
> Signed-off-by: Matthew Maurer <mmaurer@google.com>

Modulo doc suggestions from other reviews:

Reviewed-by: Alice Ryhl <aliceryhl@google.com>

On Fri, Dec 26, 2025 at 07:15:10PM +0000, Matthew Maurer wrote:
> +/// Adds the given buffer to the entropy pool, but does not credit any entropy.
> +///
> +/// This is intended for use mixing in data that is likely to differ between devices or boots, but
Yeah, I agree with Alexandre here. I think this would probably be better
written as: "This function mixes in data that is likely to differ between
devices or boots". Or, "add_device_randomness mixes in data that is
likely to differ between devices or boots." If you would prefer to avoid
saying "this function".
> +/// may otherwise be predictable. Examples include MAC addresses or RTC values. This slightly
> +/// improves randomness in entropy-constrained environments (especially common for embedded
> +/// devices).
> +pub fn add_device_randomness(buf: &[u8]) {
> +    // SAFETY: We just need the pointer to be valid for the length, which a slice provides.
> +    unsafe { bindings::add_device_randomness(buf.as_ptr().cast::<c_void>(), buf.len()) };
> +}
> 
> ---
> base-commit: 008d3547aae5bc86fac3eda317489169c3fda112
> change-id: 20251029-add-entropy-f57e12ebe110
> 
> Best regards,
> -- 
> Matthew Maurer <mmaurer@google.com>
> 
>

On Sun, Dec 28, 2025 at 1:13 AM Brendan Shephard <bshephar@bne-home.net> wrote:
>
> On Fri, Dec 26, 2025 at 07:15:10PM +0000, Matthew Maurer wrote:
> > +/// Adds the given buffer to the entropy pool, but does not credit any entropy.
> > +///
> > +/// This is intended for use mixing in data that is likely to differ between devices or boots, but
> Yeah, I agree with Alexandre here. I think this would probably be better
> written as: "This function mixes in data that is likely to differ between
> devices or boots". Or, "add_device_randomness mixes in data that is
> likely to differ between devices or boots." If you would prefer to avoid
> saying "this function".

I believe the grammar to be correct here - if it helps, the "use" in
this case is the "yoos" variant of use, not the "yooz" variant.

That said, if it's confusing, it's confusing. I'll reword it as suggested.

> > +/// may otherwise be predictable. Examples include MAC addresses or RTC values. This slightly
> > +/// improves randomness in entropy-constrained environments (especially common for embedded
> > +/// devices).
> > +pub fn add_device_randomness(buf: &[u8]) {
> > +    // SAFETY: We just need the pointer to be valid for the length, which a slice provides.
> > +    unsafe { bindings::add_device_randomness(buf.as_ptr().cast::<c_void>(), buf.len()) };
> > +}
> >
> > ---
> > base-commit: 008d3547aae5bc86fac3eda317489169c3fda112
> > change-id: 20251029-add-entropy-f57e12ebe110
> >
> > Best regards,
> > --
> > Matthew Maurer <mmaurer@google.com>
> >
> >

On Sat Dec 27, 2025 at 4:15 AM JST, Matthew Maurer wrote:
> Adds just enough support to allow device drivers to feed entropy to the
> central pool.
>
> Signed-off-by: Matthew Maurer <mmaurer@google.com>
> ---
> Changes in v3:
> - Fixed doclink to be srctree-based instead of relative.
> - Switched to prelude import instead of fine-grained imports.
> - Link to v2: https://lore.kernel.org/r/20251216-add-entropy-v2-1-4d866f251474@google.com
>
> Changes in v2:
> - Added more details in the docs about the API, specifically about it
>   not crediting entropy and when it ought to be used.
> - Link to v1: https://lore.kernel.org/r/20251212-add-entropy-v1-1-e70ad1bc9c65@google.com
> ---
>  rust/kernel/lib.rs  |  1 +
>  rust/kernel/rand.rs | 18 ++++++++++++++++++
>  2 files changed, 19 insertions(+)
>
> diff --git a/rust/kernel/lib.rs b/rust/kernel/lib.rs
> index f812cf12004286962985a068665443dc22c389a2..bf64752d276b0bdea06ac0de8a5e219190129377 100644
> --- a/rust/kernel/lib.rs
> +++ b/rust/kernel/lib.rs
> @@ -130,6 +130,7 @@
>  pub mod ptr;
>  #[cfg(CONFIG_RUST_PWM_ABSTRACTIONS)]
>  pub mod pwm;
> +pub mod rand;
>  pub mod rbtree;
>  pub mod regulator;
>  pub mod revocable;
> diff --git a/rust/kernel/rand.rs b/rust/kernel/rand.rs
> new file mode 100644
> index 0000000000000000000000000000000000000000..3e0aef2f53af122f510b34c4743a20452aaad9ad
> --- /dev/null
> +++ b/rust/kernel/rand.rs
> @@ -0,0 +1,18 @@
> +// SPDX-License-Identifier: GPL-2.0
> +
> +//! Randomness.

Let's promote this comment to a sentence at least. :)

> +//!
> +//! C header: [`include/linux/random.h`](srctree/include/linux/random.h)
> +
> +use crate::prelude::*;
> +
> +/// Adds the given buffer to the entropy pool, but does not credit any entropy.
> +///
> +/// This is intended for use mixing in data that is likely to differ between devices or boots, but

Not a native speaker, but "use mixing" sounds a bit off?

With this, and fwiw,

Reviewed-by: Alexandre Courbot <acourbot@nvidia.com>