IMHO, Kurt's suggestion sounds intriguing and definitely worth attention.
One can obtain exclussiveness by duplicating the module_list with some
sort of "functionality_list" with entry for each invocation of the init
routine of a driver (yes, hundreds of drivers will have to be modified
to call some sort of register_functionality() but who cares about the
amount of work?) be it a module or compiled-in driver.
It seems to me that the direction to think in, to solve this problem is
providing a unified way of initializing a driver, whether it is loaded as
a module via its init_module() routine or as statically linked.
Interesting problem, very interesting...
It is as if we needed another create_module() which is the same as the
usual one except the vmalloc() stuff, through which one could go even
during normal (boot) initialization sequence, registering each
"functionality" as a "built-in" module.
Regards,
------
Tigran A. Aivazian | http://www.sco.com
Escalations Research Group | tel: +44-(0)1923-813796
Santa Cruz Operation Ltd | http://www.ocston.org/~tigran
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/