From mboxrd@z Thu Jan 1 00:00:00 1970 From: Mark Salter To: nsd@redhat.com Cc: fche@redhat.com, cagney@redhat.com, jtc@redback.com, sid@sources.redhat.com, gdb-patches@cygnus.com Subject: Re: Possible remote.c patch for Z-packet breakpoints + Harvard + SID Date: Mon, 16 Jul 2001 16:42:00 -0000 Message-id: <200107162342.f6GNgUn12498@deneb.localdomain> References: <200107161931.f6GJVV719853@rtl.cygnus.com> X-SW-Source: 2001-q3/msg00016.html >>>>> Nick Duffek writes: > Executive summary for the lists I've added to this thread: Z-packet > breakpoints don't work with SID on Harvard architectures. > On 16-Jul-2001, I wrote: >> When a Harvard-architecture GDB sends an address to SID, the address has a >> high bit set indicating whether the address space, e.g. instruction >> vs. data. FWIW, I recently had to deal with this in RedBoot. Since the RedBoot stub already had to deal with differentiating code and data addresses, I just used the same mechanism to convert from a "gdb" address to a hw code or data address when decoding the Z packet before the __set_breakpoint() is called. --Mark