This is the mail archive of the
gdb-prs@sourceware.org
mailing list for the GDB project.
Re: threads/2395: "info stack" works, but "stack-info-depth" fails
- From: Daniel Jacobowitz <drow at false dot org>
- To: nobody at sources dot redhat dot com
- Cc: gdb-prs at sources dot redhat dot com,
- Date: 4 Jan 2008 12:58:02 -0000
- Subject: Re: threads/2395: "info stack" works, but "stack-info-depth" fails
- Reply-to: Daniel Jacobowitz <drow at false dot org>
The following reply was made to PR threads/2395; it has been noted by GNATS.
From: Daniel Jacobowitz <drow@false.org>
To: =?iso-8859-1?Q?=D8yvind?= Harboe <oyvind.harboe@zylin.com>
Cc: gdb-gnats@sources.redhat.com
Subject: Re: threads/2395: "info stack" works, but "stack-info-depth" fails
Date: Fri, 4 Jan 2008 07:55:00 -0500
On Fri, Jan 04, 2008 at 10:22:33AM +0100, Øyvind Harboe wrote:
> Is it realistic to have this fixed in GDB or should hardware debuggers
> simply return a garbage value
> (0xfffff...?) and let the target memory read succeed? Fixing OpenOCD
> wouldn't be hard.
It's a GDB bug. When someone has time to work on it, we'll fix it in
GDB. I would call that breaking OpenOCD, not fixing it.
> The *really* interesting bit is that GDB actually manages to figure
> out the stack depth when I simply
> removed memory_error() from corefile.c/read_memory(), i.e. GDB fails
> to read some value it does not
> actually need to evaluate the stack frames! :-)
GDB is figuring out if there is another stack frame past the top of
your stack. If you'd given it the right values, you'd have gotten
another frame in the backtrace.
--
Daniel Jacobowitz
CodeSourcery