On Fri Oct 23, 2026 at 5:40 PM CEST, Simon Song wrote:
> +    /// Prepare for runtime suspend, without holding the PCI device lock.
> +    #[cfg(CONFIG_PM)]
> +    fn runtime_suspend<'bound>(
> +        dev: &Device<device::Bound>,
> +        this: Pin<&Self::Data<'bound>>,
> +    ) -> Result
> +    where
> +        for<'a> Self::Data<'a>: Sync,
> +    {
> +        let _ = (dev, this);
> +        build_error!(crate::error::VTABLE_DEFAULT_ERROR)
> +    }
> +
> +    /// Finish runtime resume, without holding the PCI device lock.
> +    #[cfg(CONFIG_PM)]
> +    fn runtime_resume<'bound>(dev: &Device<device::Bound>, this: 
> Pin<&Self::Data<'bound>>) -> Result
> +    where
> +        for<'a> Self::Data<'a>: Sync,
> +    {
> +        let _ = (dev, this);
> +        build_error!(crate::error::VTABLE_DEFAULT_ERROR)
> +    }
> +
> +    /// Runtime PM callbacks, if any. Construct with [`PmOps::new`].
> +    #[cfg(CONFIG_PM)]
> +    const PM_OPS: Option<PmOps<Self>> = None;

As mentioned in my reply to patch 6, please use the generic runtime PM code.

Reply via email to