On 01/09/11 10:15, Manohar Vanga wrote:
> diff --git a/drivers/staging/vme/vme.h b/drivers/staging/vme/vme.h
> index 4155d8c..d442cce 100644
> --- a/drivers/staging/vme/vme.h
> +++ b/drivers/staging/vme/vme.h
> @@ -91,17 +91,34 @@ extern struct bus_type vme_bus_type;
> #define VME_SLOT_CURRENT -1
> #define VME_SLOT_ALL -2
>
> +/**
> + * VME device identifier structure
> + * @bus: The bus ID of the bus the device is on
> + * @slot: The slot this device is plugged into
> + */
> struct vme_device_id {
> int bus;
> int slot;
> };
>
> +/**
> + * Structure representing a VME device
> + * @id: The ID of the device (currently the bus and slot number)
> + * @bridge: Pointer to the bridge device this device is on
> + * @dev: Internal device structure
> + */
> +struct vme_dev {
> + struct vme_device_id id;
> + struct vme_bridge *bridge;
> + struct device dev;
> +};
> +
I think we can probably merge vme_device_id and vme_dev.
Since we have a pointer to the vme_bridge, the bus number in vme_device_id is
kinda superfluous.
The direction we are heading in makes the slot number far less important, in
some ways it becomes more of an optional information field (in pre-vme64 racks
we probably won't know which slot the device we are bound to is in anyway).
Basically moving from a binding mechanism like PCI to a binding mechanism
closer to ISA, which matches the (widely used, more historic subset) of
capabilites of the VME bus. I think we can just move the slot number to
vme_dev and do away with vme_device_id entirely.
Martyn
--
Martyn Welch (Principal Software Engineer) | Registered in England and
GE Intelligent Platforms | Wales (3828642) at 100
T +44(0)1327322748 | Barbirolli Square, Manchester,
E martyn.welch@xxxxxx | M2 3AB VAT:GB 927559189
_______________________________________________
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxx
http://driverdev.linuxdriverproject.org/mailman/listinfo/devel
