Re: [RFC PATCH 2/2] libxl: prototype libxl_device_nic_add/remove with IDL

[Anthony PERARD <anthony.perard@xxxxxxxxxx>]

On Mon, Jul 27, 2020 at 09:26:33AM -0400, Nick Rosbrook wrote:
> Add a DeviceFunction class and describe prototypes for
> libxl_device_nic_add/remove in libxl_types.idl.
> 
> Signed-off-by: Nick Rosbrook <rosbrookn@xxxxxxxxxxxx>
> --
> This is mostly to serve as an example of how the first patch would be
> used for function support in the IDL.
> ---
>  tools/libxl/idl.py          | 8 ++++++++
>  tools/libxl/libxl_types.idl | 6 ++++++
>  2 files changed, 14 insertions(+)
> 
> diff --git a/tools/libxl/idl.py b/tools/libxl/idl.py
> index 1839871f86..15085af8c7 100644
> --- a/tools/libxl/idl.py
> +++ b/tools/libxl/idl.py
> @@ -386,6 +386,14 @@ class CtxFunction(Function):
>  
>          Function.__init__(self, name, params, return_type, return_is_status)
>  
> +class DeviceFunction(CtxFunction):
> +    """ A function that modifies a device. """

I guess that meant to be used by all function generated with the C macro
LIBXL_DEFINE_DEVICE_ADD() and LIBXL_DEFINE_DEVICE_REMOVE(), isn't it?

I wonder if if we could get away with the type of device ("nic") and the
type of the parameter (`libxl_device_nic`) and have DeviceFunction been
a generator for both `add` and `remove` functions (and `destroy`).

Also there are functions like libxl_devid_to_device_nic() aren't those
of type DeviceFunction as well ? But they don't takes any `ao_how`.

There is also `libxl_device_nic_list{,_free}`, but it is to handle a
list of libxl_device_*, so it could be kind of related to DeviceFunction, but
not quite. But maybe I'm going to far :-).

> diff --git a/tools/libxl/libxl_types.idl b/tools/libxl/libxl_types.idl
> +libxl_device_nic_add = DeviceFunction("device_nic_add",
> +    ("nic", libxl_device_nic))
> +
> +libxl_device_nic_remove = DeviceFunction("device_nic_remove",
> +    ("nic", libxl_device_nic))
> +

Thanks,

-- 
Anthony PERARD