Re: [RFC][PATCH] input: Introduce device information ioctl

[Kay Sievers]

On Tue, Dec 7, 2010 at 13:44, Arnd Bergmann <arnd@xxxxxxxx> wrote:
> On Tuesday 07 December 2010, Kay Sievers wrote:
>> On Tue, Dec 7, 2010 at 11:56, Dmitry Torokhov <dmitry.torokhov@xxxxxxxxx> wrote:
>> > On Tue, Dec 07, 2010 at 11:48:28AM +0100, Kay Sievers wrote:
>>
>> >> Please don't add new ioctls which are not extensible. The ioctl should
>> >> carry the length or the version of the structure it asks for, so it
>> >> can be extended in the future.
>> >
>> > Size of ioctl data is encoded in ioctl, it can be extended easily. For
>> > examples take a look at how EVIOCGKEYCODE and EVIOCGSKEYCODE are handled
>> > in recent kernels.
>>
>> Oh, how does that work? With the ioctl call, userspace has to supply
>> the size it expects to be returned from the kernel. How does the
>> kernel otherwise know how much it is allowed to copy to the user?
>
> The ioctl command number itself is calculated from the size of the
> data that gets passed:
>
> #define EVIOCGDEVINFO Â Â Â Â Â_IOR('E', 0x09, struct input_devinfo)
>
> If struct input_devinfo ever changes (which it can, but should not),
> the command changes as well.

So unlike statet, it's not extensible, and this struct and this ioctl
can never change?

Kay
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/