On Monday 26 October 2009, Øyvind Harboe wrote: > I.e. support for software breakpoints to MMU read only memory > must be implemented in target_write_memory() directly. Unless > we expose cache flushing through a polymorphic interface(shudder...)
The core will need to become somewhat MMU-aware; I don't see any way around that eventuality. Knowing that the mapping is read-only is one thing. Cache flushing will need to become moderately generic. And possibly even understanding about *shared* read-only mappings will matter. As in, if the mapping is shared, other problems come up... _______________________________________________ Openocd-development mailing list [email protected] https://lists.berlios.de/mailman/listinfo/openocd-development
