Maurizio Lombardi wrote:
> I'm wondering about this "standard interface", do you think it is correct to
> add
> a new clock_dev_ops_t interface that every rtc driver must implement?
> I'm talking about something like the following:
>
> typedef struct {
> int (*time_set)(ddf_fun_t *, time_t);
> int (*time_get)(ddf_fun_t *, time_t *);
> } clock_dev_ops_t;
>
> Am I on the right track?
I guess so. Are you asking about the interface itself (i.e. the choice of
methods and their signatures)? Looks reasonable, though only when there are
more drivers we will see for certain whether the design will need to adapt :-)
Or are you asking about the form of the interface (clock_dev_ops_t)? As you
should know you have currently a choice in driver either the use the DDF's
dispatcher or write all the IPC code by hand just like for a regular server.
Unless you plan for a non-DDF server that would implement the same interface,
it probably doesn't matter much - and changing from one to the other is not
extremely difficult. This is a choice that will be eliminated in the future -
when ticket #424 (RPC/IPC generator) is implemented, there will no longer be a
need for DDF's specialized dispatcher.
Regards
-Jiri
_______________________________________________
HelenOS-devel mailing list
[email protected]
http://lists.modry.cz/cgi-bin/listinfo/helenos-devel
