This is the mail archive of the
gdb-prs@sources.redhat.com
mailing list for the GDB project.
Re: gdb/47: Single-stepping on IA-64 ignores register rotation
- To: nobody at sources dot redhat dot com
- Subject: Re: gdb/47: Single-stepping on IA-64 ignores register rotation
- From: Kevin Buettner <kevinb at cygnus dot com>
- Date: 27 Apr 2001 23:18:00 -0000
- Cc: gdb-prs at sources dot redhat dot com,
- Reply-To: Kevin Buettner <kevinb at cygnus dot com>
The following reply was made to PR gdb/47; it has been noted by GNATS.
From: Kevin Buettner <kevinb@cygnus.com>
To: bob@virtualunlimited.com, gdb-gnats@sources.redhat.com
Cc:
Subject: Re: gdb/47: Single-stepping on IA-64 ignores register rotation
Date: Fri, 27 Apr 2001 16:13:16 -0700
On Apr 26, 9:31am, bob@virtualunlimited.com wrote:
> >Synopsis: Single-stepping on IA-64 ignores register rotation
> >Release: gdb-5.0
> >Environment:
> ia64-suse-linux
> >Description:
> When trying to piece of vectorized code (with rotating registers and
> br.ctop as loop instruction) I noticed that single-stepping through
> the code ignores register rotation, i.e. stepping through the
> br.ctop instruction changes neither regular registers, predicate
> registers nor floating-point registers. Of course this makes it
> virtually impossible to single-step and debug that piece of code.
> >How-To-Repeat:
> Write any loop with br.ctop, for instance loading register r32 from
> an increasing address. Single-step through the routine, and see how
> the registers haven't rotated after each br.ctop.
Bob,
It isn't clear to me which version of gdb Suse is including in their
IA-64 Linux distribution. In any event, I would appreciate it if you'd
try again with the current development sources since I implemented
register rotation support quite some time ago. If it still doesn't
work for you, I'd appreciate it if you'd send me a test case to look
at.
Thanks,
Kevin