phcoder wrote: > Well the interface is as we described: the module gives a callback > function which will be called before launching boot function. This > interface is enough for both (and probaly many other) needs. The only > problem is that callback functions can conflict with each other and with > boot function. E.g. if callback sets an INT13h hook and then boot > function reads harddrive then it could get wrong data. In my opinion the > ese callback and boot functions shouldn't use device access at all. This > is especially true because theese functions are after grub_machine_fini.
That is more like a textual description of the issues not an interface description. It has to be more concrete than that. And mechanism should support hook installation. And I do not see reason why that same thing cannot support also this send key stuff. >> This way it would be also easier to incorporate patches as there is >> already skeleton that can be used easily. >> > There could several templates for using such feature because it's quite > universal I am not really delighted about that... Just one or two connection points and that should be able to handle those cases. If not then we have something wrong somewhere... _______________________________________________ Grub-devel mailing list Grub-devel@gnu.org http://lists.gnu.org/mailman/listinfo/grub-devel
