rust/kernel/alloc/kvec/errors.rs | 3 +++ rust/kernel/error.rs | 6 ++++++ rust/kernel/ptr/projection.rs | 2 +- rust/kernel/xarray.rs | 1 + rust/syn/error.rs | 1 + 5 files changed, 12 insertions(+), 1 deletion(-)
From: Alistair Francis <alistair.francis@wdc.com>
Mark all of the existing
impl From<...> for Error {
fn from(err: ...) -> Self {
...
}
}
functions as `#[inline]`
There was a recent request [1] to inline the simple from() Error
functions to avoid the function call overhead.
This patch inlines all of the functions (except for the pin-init) and
converts one existing `inline(always)` to `inline` to match the styles.
1: https://lkml.org/lkml/2026/3/13/962
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
---
The pin-init changes have been submitted here instead:
https://github.com/Rust-for-Linux/pin-init/pull/126
rust/kernel/alloc/kvec/errors.rs | 3 +++
rust/kernel/error.rs | 6 ++++++
rust/kernel/ptr/projection.rs | 2 +-
rust/kernel/xarray.rs | 1 +
rust/syn/error.rs | 1 +
5 files changed, 12 insertions(+), 1 deletion(-)
diff --git a/rust/kernel/alloc/kvec/errors.rs b/rust/kernel/alloc/kvec/errors.rs
index e7de5049ee47..985c5f2c3962 100644
--- a/rust/kernel/alloc/kvec/errors.rs
+++ b/rust/kernel/alloc/kvec/errors.rs
@@ -15,6 +15,7 @@ fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
}
impl<T> From<PushError<T>> for Error {
+ #[inline]
fn from(_: PushError<T>) -> Error {
// Returning ENOMEM isn't appropriate because the system is not out of memory. The vector
// is just full and we are refusing to resize it.
@@ -32,6 +33,7 @@ fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
}
impl From<RemoveError> for Error {
+ #[inline]
fn from(_: RemoveError) -> Error {
EINVAL
}
@@ -55,6 +57,7 @@ fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
}
impl<T> From<InsertError<T>> for Error {
+ #[inline]
fn from(_: InsertError<T>) -> Error {
EINVAL
}
diff --git a/rust/kernel/error.rs b/rust/kernel/error.rs
index 258b12afdcba..935787c2a91c 100644
--- a/rust/kernel/error.rs
+++ b/rust/kernel/error.rs
@@ -216,36 +216,42 @@ fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
}
impl From<AllocError> for Error {
+ #[inline]
fn from(_: AllocError) -> Error {
code::ENOMEM
}
}
impl From<TryFromIntError> for Error {
+ #[inline]
fn from(_: TryFromIntError) -> Error {
code::EINVAL
}
}
impl From<Utf8Error> for Error {
+ #[inline]
fn from(_: Utf8Error) -> Error {
code::EINVAL
}
}
impl From<LayoutError> for Error {
+ #[inline]
fn from(_: LayoutError) -> Error {
code::ENOMEM
}
}
impl From<fmt::Error> for Error {
+ #[inline]
fn from(_: fmt::Error) -> Error {
code::EINVAL
}
}
impl From<core::convert::Infallible> for Error {
+ #[inline]
fn from(e: core::convert::Infallible) -> Error {
match e {}
}
diff --git a/rust/kernel/ptr/projection.rs b/rust/kernel/ptr/projection.rs
index 140ea8e21617..7e37390da6c6 100644
--- a/rust/kernel/ptr/projection.rs
+++ b/rust/kernel/ptr/projection.rs
@@ -13,7 +13,7 @@
pub struct OutOfBound;
impl From<OutOfBound> for Error {
- #[inline(always)]
+ #[inline]
fn from(_: OutOfBound) -> Self {
ERANGE
}
diff --git a/rust/kernel/xarray.rs b/rust/kernel/xarray.rs
index a49d6db28845..46e5f43223fe 100644
--- a/rust/kernel/xarray.rs
+++ b/rust/kernel/xarray.rs
@@ -172,6 +172,7 @@ pub struct StoreError<T> {
}
impl<T> From<StoreError<T>> for Error {
+ #[inline]
fn from(value: StoreError<T>) -> Self {
value.error
}
diff --git a/rust/syn/error.rs b/rust/syn/error.rs
index 6fa0faf7f4e4..636f07d56d05 100644
--- a/rust/syn/error.rs
+++ b/rust/syn/error.rs
@@ -405,6 +405,7 @@ impl Copy for SpanRange {}
impl std::error::Error for Error {}
impl From<LexError> for Error {
+ #[inline]
fn from(err: LexError) -> Self {
Error::new(err.span(), err)
}
--
2.53.0On Wed, Mar 25, 2026 at 1:51 AM <alistair23@gmail.com> wrote: > > 1: https://lkml.org/lkml/2026/3/13/962 For links, please use the Link: tag (and also please use lore.kernel.org for mailing list links), e.g. it would usually look like: Link: https://lore.kernel.org/... [1] (By the way, the idea was to create a "good first issue" for new contributors in our issue tracker, but of course it is also good to get this done) > rust/kernel/alloc/kvec/errors.rs | 3 +++ > rust/kernel/xarray.rs | 1 + Thanks for Cc'ing the maintainers on this -- now the maintainers may give their Acked-by's! :) > rust/kernel/ptr/projection.rs | 2 +- Gary: there is no `build_assert!` in sight, so it is fine, but perhaps you wanted to have it more forcefully inlined because it was trivial? > rust/syn/error.rs | 1 + We don't want to change `syn`, since it is a vendored crate. It also runs just in the host, so it doesn't matter if it is inline or not. Thanks! Cheers, Miguel
On Wed Mar 25, 2026 at 2:19 AM CET, Miguel Ojeda wrote: >> rust/kernel/alloc/kvec/errors.rs | 3 +++ > > Thanks for Cc'ing the maintainers on this -- now the maintainers may > give their Acked-by's! :) Acked-by: Danilo Krummrich <dakr@kernel.org>
On Wed Mar 25, 2026 at 1:19 AM GMT, Miguel Ojeda wrote: > On Wed, Mar 25, 2026 at 1:51 AM <alistair23@gmail.com> wrote: >> >> 1: https://lkml.org/lkml/2026/3/13/962 > > For links, please use the Link: tag (and also please use > lore.kernel.org for mailing list links), e.g. it would usually look > like: > > Link: https://lore.kernel.org/... [1] > > (By the way, the idea was to create a "good first issue" for new > contributors in our issue tracker, but of course it is also good to > get this done) > >> rust/kernel/alloc/kvec/errors.rs | 3 +++ >> rust/kernel/xarray.rs | 1 + > > Thanks for Cc'ing the maintainers on this -- now the maintainers may > give their Acked-by's! :) > >> rust/kernel/ptr/projection.rs | 2 +- > > Gary: there is no `build_assert!` in sight, so it is fine, but perhaps > you wanted to have it more forcefully inlined because it was trivial? I just `#[inline(always)]` everything there :) I don't think this path is actually exercised by `build_assert!()`, so it is fine. > >> rust/syn/error.rs | 1 + > > We don't want to change `syn`, since it is a vendored crate. It also > runs just in the host, so it doesn't matter if it is inline or not. I wouldn't mind having it, but it should go separately and not be part of this patch. Best, Gary
© 2016 - 2026 Red Hat, Inc.