>> > I'm fine with the change from "const void" to "const char", but we need >> > to remove a preceding comment about void labels. >> It's not that I'm opposed to void in principle. Just using the same >> constructions to do the same things in different files makes code >> easier to learn and port > > I wonder if we can go the other way and use void for all labels without > storage. Indeed, it's too easy to misuse a char variable by forgetting > the ampersand before it. If you define it as char [] (as done everywhere in grub2), not char, you need no ampersand. > >> > As for the parse_biosdisk() change, I'd like to see an explanation. >> The explanation is that if user uses ata or usbms code and code calls >> biosdisk, BIOS may issue a command which may conflict with ata/usbms. >> Unfortunately it's not a scenario we're able to circumvent (BIOS is >> headache) so I prefer to err on a safe side > > I agree that we should avoid touching the hardware. Besides, after > loading ata we may not see some drives that BIOS can see. > > Validation of user input is good, but only if it's implemented > correctly. > > Another approach may be to use biosdisk calls only if biosdisk is > active. Otherwise, trust the user and turn off validation. > Not really feasible since with our design "active" driver isn't well defined. > -- > Regards, > Pavel Roskin > > > _______________________________________________ > Grub-devel mailing list > Grub-devel@gnu.org > http://lists.gnu.org/mailman/listinfo/grub-devel >
-- Regards Vladimir 'phcoder' Serbinenko _______________________________________________ Grub-devel mailing list Grub-devel@gnu.org http://lists.gnu.org/mailman/listinfo/grub-devel
