Re: [PATCH v4 05/10] vpci: Hide legacy capability when it fails to initialize

[Jan Beulich <jbeulich@xxxxxxxx>]

On 09.05.2025 11:05, Jiqian Chen wrote:
> @@ -83,6 +99,100 @@ static int assign_virtual_sbdf(struct pci_dev *pdev)
>  
>  #endif /* CONFIG_HAS_VPCI_GUEST_SUPPORT */
>  
> +static struct vpci_register *vpci_get_register(struct vpci *vpci,
> +                                               unsigned int offset,
> +                                               unsigned int size)
> +{
> +    const struct vpci_register r = { .offset = offset, .size = size };
> +    struct vpci_register *rm;
> +
> +    ASSERT(spin_is_locked(&vpci->lock));
> +    list_for_each_entry ( rm, &vpci->handlers, node )
> +    {
> +        int cmp = vpci_register_cmp(&r, rm);
> +
> +        if ( !cmp && rm->offset == offset && rm->size == size )

What's the point of using vpci_register_cmp() when you need to do
the "exact match" check here anyway?

> +static int vpci_capability_mask(struct pci_dev *pdev, unsigned int cap)

What's the word "mask" indicating here? The function doesn't return any
mask afaics. Do you perhaps mean "hide"?

Jan