On Tuesday 19 January 2010, Øyvind Harboe wrote: > >> With the str710 we have something nice to test against since the > >> there are at least three sector sizes and two banks. Pretty messy :-) > > > > You sent the following memory map: > > > >> (gdb) info mem > >> 0 y 0x00000000 0x40000000 rw nocache > >> 1 y 0x40000000 0x40040000 flash blocksize 0x2000 nocache > >> 2 y 0x40040000 0x400c0000 rw nocache > >> 3 y 0x400c0000 0x400c4000 flash blocksize 0x2000 nocache > >> 4 y 0x400c4000 0x100000000 rw nocache > > > > That shows two banks, yes ... but only one sector size, not three. > > That's because the memory map is incorrectly generated in > gdb_server.c. > > gdb_server.c needs to iterate over banks & then sectors within > each bank, rather than assuming that all sectors in a bank > are of equal size.
Right. In preparation for that, the second pach I just sent. > > SEPARATE ISSUE ... I think we should consider (for 0.5) adding > > ways to declare where RAM/SRAM regions exist. Both from C code > > (microcontrollers can often autoprobe that stuff) and Tcl. It's > > unwise to just [guess] anything about non-flash regions. > > That would be an excellent refinement. Also we can "cache" such > regions in order to improve gdb performance, another refinement. Not sure what you mean by caching ... if the CPU is running, we can't assume it's not going to touch such areas. Though: I looked at the GDB protocol spec and it says that undefined areas are presumed to be RAM. So I'm a bit puzzled about just what that current code is there for... - Dave _______________________________________________ Openocd-development mailing list [email protected] https://lists.berlios.de/mailman/listinfo/openocd-development
