On Mon, Mar 02, 2026 at 04:01:44PM +0100, David Hildenbrand (Arm) wrote:

On 3/2/26 11:33, Alice Ryhl wrote:

quoted

On Mon, Mar 02, 2026 at 11:27:40AM +0100, David Hildenbrand (Arm) wrote:

quoted

On 3/2/26 11:01, Alice Ryhl wrote:

quoted

Well, rustfmt comes with the compiler, and it would be ideal to build
test changes before sending them :)

At least on Ubuntu on my notebook where I do most of the coding+patch
submissions it's a separate package?

I do all my builds on a different (more powerful) machine where the
whole rust machinery's in place. Further, build bots that run on my
private branches did not report any issues.

There are some build bots that check for rustfmt, though not all of
them.

quoted

quoted

But no worries, I took care of testing it. Thanks for taking the time to
update the Rust code as well.

I just did an allyesconfig and it does not report any warnings.

So apparently, rustfmt problems not result in the compiler complaining?

Or something else is off here that rust/kernel/mm/virt.rs won't get
compiled on my machine, even with allyesconfig. I can definitely see
some RUSTC stuff happening in the logs, like

	RUSTC L rust/kernel.o

Thanks for the review and for pointing out rustfmt!

Similar to kerneldoc and other similar targets, formatting isn't checked
in the normal build, but make can be invoked on the rustfmtcheck target
to check it.

Thanks adding that to my cross-compile chain.

Awesome, thanks!

It's not relevant in this patch, but another thing that may be useful is
to add CLIPPY=1 to the make invocation when building normally. This
causes additional warnings to be checked using a tool called clippy.

Alice