This is the mail archive of the
gdb@sources.redhat.com
mailing list for the GDB project.
Re: question software_single_step / breakpoint_from_pc
- To: Michael Staudenmaier <Michael dot Staudenmaier at motorola dot com>
- Subject: Re: question software_single_step / breakpoint_from_pc
- From: Andrew Cagney <ac131313 at cygnus dot com>
- Date: Thu, 20 Sep 2001 11:48:35 -0400
- Cc: gdb at sources dot redhat dot com
- References: <3BA9C547.A1CD4688@motorola.com>
Yes, you've hit a known limitation in the way the current software
single step interface is implemented. Look for single_step and fixme in
gdbarch.*. Patches definitly welcome.
Andrew
> Hi,
> porting gdb to a new target platform i faced the following problem:
>
> Scenario:
> Our target processor itself does not support single step & hardware
> breakpoints.
> But there are two "flavors" of the target system
>
> 1. simulation which supports hardware breakpoints & single step (feature
> of RTL simulator)
>
> 2. concrete hardware using a debugger in ROM via serial link. (no
> hardware bp & no single step)
>
> I wrote one target description file (corename-tdep.c) and two remote_*.c
> driver for the versions mentioned.
>
> Question:
> Currently i can only decide that
> - software single step
> - hardware breakpoints
> are available or not.
> (decr_pc_after_break also varies depending on hw-breakpoint
> availibility)
>
> How can i make this dependend on the target connection (remote_*.c)
> driver selected?
>
> Your answers are greatly appreciated!
>
> Regards,
>
> Michael
> --
>