Re: [PATCH v11 2/7] rust: debugfs: Add support for read-only files

[Greg Kroah-Hartman <gregkh@linuxfoundation.org>]

On Mon, Sep 08, 2025 at 03:22:41PM +0200, Danilo Krummrich wrote:
> On Mon Sep 8, 2025 at 2:48 PM CEST, Greg Kroah-Hartman wrote:
> > On Mon, Sep 08, 2025 at 12:54:46PM +0200, Danilo Krummrich wrote:
> >> diff --git a/samples/rust/rust_debugfs.rs b/samples/rust/rust_debugfs.rs
> >> index b26eea3ee723..475502f30b1a 100644
> >> --- a/samples/rust/rust_debugfs.rs
> >> +++ b/samples/rust/rust_debugfs.rs
> >> @@ -59,6 +59,8 @@ struct RustDebugFs {
> >>      #[pin]
> >>      _compatible: File<CString>,
> >>      #[pin]
> >> +    _test: File<&'static CStr>,
> >> +    #[pin]
> >>      counter: File<AtomicUsize>,
> >>      #[pin]
> >>      inner: File<Mutex<Inner>>,
> >> @@ -140,6 +142,7 @@ fn new(pdev: &platform::Device<Core>) -> impl PinInit<Self, Error> + '_ {
> >>                          .property_read::<CString>(c_str!("compatible"))
> >>                          .required_by(dev)?,
> >>                  ),
> >> +                _test <- debugfs.read_only_file(c_str!("test"), c_str!("some_value")),
> >
> > Cool, but again, we do not want to ever be storing individual debugfs
> > files.  Well, we can, but for 90% of the cases, we do not, we only want
> > to remove the whole directory when that goes out of scope, which will
> > clean up the files then.
> 
> This API does not work in the way that you have a struct storing the data you
> want to expose *and* another one for the files with the data attached.
> 
> The File type contains the actual data. For instance, if you have a struct Foo,
> where you want to expose the members through debugfs you would *not* do:
> 
> 	struct Foo {
> 	   a: u32,
> 	   b: u32,
> 	}
> 
> 	struct FooFiles {
> 	   a: File<&u32>,
> 	   b: File<&u32>
> 	}
> 
> and then create an instance of Foo *and* another instance of FooFiles to export
> them via debugfs.

Ah, that's exactly what I was trying to do.

> Instead you would change your struct Foo to just be:
> 
> 	struct Foo {
> 	   a: File<u32>,
> 	   b: File<u32>,
> 	}
> 
> If you now create an instance of Foo (let's call it `foo`), then foo.a or foo.b
> dereferences to the inner type, i.e. the u32. Or in other words `foo` still
> behaves as if `a` and `b` would be u32 values. For instance:
> 
>    if foo.a == 42 {
>       pr_info!("Foo::b = {}\n", foo.b);
>    }

Oh that's not going to work well at all :(

Think about something "simple" like a pci config descriptor.  You have a
structure, with fields, already sitting there.  You want to expose those
fields in debugfs.  So you want to only create debugfs files in one
location in a driver, you don't want ALL users of those fields to have
to go through a File<T> api, right?  That would be crazy, all drivers
would end up always having File<T> everywhere.

> The fact that the backing files of `a` and `b` are removed from debugfs when Foo
> is dropped is necessary since otherwise we create a UAF.

That's fine, but:

> Think of File<T> as a containers like you think of KBox<T>.

Ok, but again, you are now forcing all users to think of debugfs as the
main "interface" to those variables, which is not true (nor should it
be.)

> KBox<T> behaves exactly like T, but silently manages the backing kmalloc()
> allocation that T lives in.
> 
> With File<T> it's exactly the same, it behaves exactly like the T that lives
> within File<T>, but silently manages the debugfs file the T is exposed by.

And what happens if debugfs is not enabled?  What about if creating the
file fails?  The variable still needs to be present and active and
working.

thanks,

greg k-h