Hi Maurizio, > jermar : it also made me think about what to do if there are > more devices in one category, such as clock > mlombardi : Yes I had the same doubt, I've no made any decisions yet > jermar : this is rather a generic problem > jermar : for instance, the category can have a user-defined default > jermar : and the user could be given the power to change the default > jermar : in case of some devices, depending on their use, > the user may specify the driver to use in a given context > jermar : e.g. nic > mlombardi : yes I understand > mlombardi : I think I should ask to Jiri what he thinks about it > jermar : good idea > --------------------------------- I understand what you mean, but let's approach that problem when it actually appears. I am not sure whether it will be so widespread as to require some solution at location service level. Basically until we face the problem it makes no sense trying to solve it. Your example with default NIC makes no sense, there is no such thing. A default printer would be a good example, perhaps.
> I've another request also: > > It would be nice if the loc_service_get_name() function would return a > service name with a format compatible with what the > devman_fun_get_handle() function expects. > > I mean, loc_service_get_name() returns something like this: > > devices/\hw\pci0\00:01.0\cmos-rtc\a > > but the devman_fun_get_handle() function requires a string with the > following format: > > /hw/pci0/00:01.0/cmos-rtc/a Sorry, but if you are trying to make calls to devman, you are heading in the wrong direction. Nobody except libdrv and devinfo talks to Devman. All clients are only aware of services, which are made accessible via Location Service. Clients are not aware if these services are implemented in a driver or not. Cheers -Jiri _______________________________________________ HelenOS-devel mailing list [email protected] http://lists.modry.cz/cgi-bin/listinfo/helenos-devel
