gdb-5.3: how to avoid soft breakpoints?

gdb-5.3: how to avoid soft breakpoints?

[Heiko Panther]

I have an embedded target running gdb stub on an openrisc processor 
(www.opencores.org). The application is running from ROM, so I cannot 
set soft breakpoints. I worked the gdb stub so it will ERR Z0 packets, 
and use Z[123] packets. This works perfectly well with hbreak. break 
will attempt to set a soft breakpoint and fail. So do "next" and 
"finish" when they try to set a breakpoint.

Is there any way around that? Can I have gdb totally avoid soft breakpoints?

Heiko

RE: gdb-5.3: how to avoid soft breakpoints?

[Jon Beniston]

> 
> I have an embedded target running gdb stub on an openrisc processor 
> (www.opencores.org). The application is running from ROM, so I cannot 
> set soft breakpoints. I worked the gdb stub so it will ERR Z0 
> packets, 
> and use Z[123] packets. This works perfectly well with hbreak. break 
> will attempt to set a soft breakpoint and fail. So do "next" and 
> "finish" when they try to set a breakpoint.
> 
> Is there any way around that? Can I have gdb totally avoid 
> soft breakpoints?

I recently posted a patch that tries to use h/w breakpoints when s/w
breakpoints fail. It worked for me in pretty much the same situation as
you seem to be in, but I didn't get any feedback on it. See:

http://sources.redhat.com/ml/gdb-patches/2004-01/msg00343.html

Cheers,
JonB