Re: [Openocd-development] cortex breakpoint trouble, R546

[Ø. H. <oyv...@zy...>]

On Tue, Apr 8, 2008 at 11:03 PM, Edwin Olson <eo...@mi...> wrote:
> Øyvind Harboe wrote:
>
> >
> > >  It seems that a four-byte length breakpoint is requested, but only two
> > >  byte breakpoints are supported by the cortex target. Any suggestions?
> > >
> > >
> >
> > Which version of GDB are you using and how was it built?
> >
> >
>  The "Lite" version of CodeSourcery, 6.6.50.20070821-cvs, which was
> configured with --target=arm-none-eabi.
>
>
>
> > You can also try disabling software breakpoints. Remember to remove
> > other breakpoints first so you don't run out of hw breakpoints for
> > evaluating the expression.
> >
> >
>  My reading of the cortex target code seems to indicate that it will
> automatically use hardware breakpoints whenever the breakpoint lives in
> flash (which sounds right to me). So I believe the request for a software
> breakpoint (Z0) is automatically being converted into a hardware breakpoint.
>
>  I don't understand what they mean by the "length" of a breakpoint. Is that
> the length of the instruction "patch"?

At a guess it looks like GDB is asking for a 4 byte breakpoint (the length
of a break instruction presumably).

Perhaps you need a GDB w/"proper" Cortex support?

I haven't looked into GDB + Cortex yet.

-- 
Øyvind Harboe
http://www.zylin.com - eCos ARM & FPGA developer kit