[DGD] nomask and static vs straight

[Shentino]

Sorry for rambling but...

What, generally, if anything, would distinguish the call to my own lfun
create() as opposed to the klib driver's call to, say, initd::reboot() ?

On Thu, Jan 29, 2009 at 7:26 AM, Shentino <shentino at gmail.com> wrote:

> In particular, some of these qualify as intrinsic "applies", such as the
> destruct() lfun.
>
> Closest analogy is klib's create() function.
>
> On Thu, Jan 29, 2009 at 7:00 AM, Shentino <shentino at gmail.com> wrote:
>
>>
>>
>> On Thu, Jan 29, 2009 at 6:49 AM, Felix A. Croes <felix at dworkin.nl> wrote:
>>
>>> Shentino <shentino at gmail.com> wrote:
>>>
>>> >[...]
>>> > What I'm curious about is who should bear the burden of access control.
>>> >  With a nomask and a stack, my role inheritable guarantees that a
>>> simple
>>> > static function supplied by the inheritor will suffice, whereas if I
>>> just do
>>> > a straiht call, the inheritor has to bear the burden of allowing the
>>> > appropriate caller, and yet turn away from everyone else.
>>> >
>>> > Any reason I should do one over the other?
>>>
>>> It could be a matter of responsibility and security.  The _F_hook setup
>>> may protect the consistency of data that cannot be accessed directly by
>>> the inheritor, or that could even be messed up by a maliciously-written
>>> inheritor.
>>>
>>> When the responsibility lies with the inheritor, let the inheritor do
>>> the checking.
>>>
>> Then I guess it would be do direct and let the inheritor worry about
>> security.  None of my cases involve data, rather, the daemons are calling
>> _F_hook so that the hook() defined by the inheritor can enjoy the benefits
>> of static declaration.
>>
>>> Regards,
>>> Felix Croes
>>> ___________________________________________
>>> https://mail.dworkin.nl/mailman/listinfo/dgd
>>>
>>
>>
>