Re: [PATCH v3 3/5] samples: rust: Provide example using the new Rust MiscDevice abstraction

[Lee Jones <lee@kernel.org>]

On Fri, 06 Dec 2024, Greg Kroah-Hartman wrote:

> On Fri, Dec 06, 2024 at 07:52:46AM +0100, Arnd Bergmann wrote:
> > On Fri, Dec 6, 2024, at 07:49, Greg KH wrote:
> > > On Thu, Dec 05, 2024 at 04:25:20PM +0000, Lee Jones wrote:
> > >> +
> > >> +#[vtable]
> > >> +impl MiscDevice for RustMiscDevice {
> > >> +    type Ptr = KBox<Self>;
> > >> +
> > >> +    fn open() -> Result<KBox<Self>> {
> > >> +        pr_info!("Opening Rust Misc Device Sample\n");
> > >> +
> > >> +        Ok(KBox::new(RustMiscDevice, GFP_KERNEL)?)
> > >> +    }
> > >> +
> > >> +    fn ioctl(
> > >> +        _device: <Self::Ptr as kernel::types::ForeignOwnable>::Borrowed<'_>,
> > >> +        cmd: u32,
> > >> +        _arg: usize,
> > >> +    ) -> Result<isize> {
> > >> +        pr_info!("IOCTLing Rust Misc Device Sample\n");
> > >> +
> > >> +        match cmd {
> > >> +            RUST_MISC_DEV_HELLO => pr_info!("Hello from the Rust Misc Device\n"),
> > >> +            _ => {
> > >> +                pr_err!("IOCTL not recognised: {}\n", cmd);
> > >> +                return Err(ENOIOCTLCMD);
> > >
> > > To quote errno.h:
> > > 	These should never be seen by user programs
> > >
> > > The correct value here is ENOTTY.
> > >
> > > Yeah, fun stuff.  Not intuitive at all, sorry.
> > 
> > That should get handled by vfs_ioctl() converting the
> > ENOIOCTLCMD to ENOTTY.
> 
> Ah, I always miss that, for some reason I thought that didn't happen
> there, but happened in subsystems that did ENOIOCTLCMD for their
> sub-ioctl handlers.  And yet it's always been that way, nevermind...
> 
> Anyway, I always recommend just using ENOTTY to be "safe", the usual way
> people get this wrong is using EINVAL.

No matter.  I can change it.

-- 
Lee Jones [李琼斯]