This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [rfc/rfa v2] Use ARM exception tables as GDB unwinder
- From: Daniel Jacobowitz <dan at codesourcery dot com>
- To: Ulrich Weigand <uweigand at de dot ibm dot com>
- Cc: gdb-patches at sourceware dot org, rearnsha at arm dot com, matthew dot gretton-dann at arm dot com
- Date: Thu, 9 Dec 2010 09:24:39 -0500
- Subject: Re: [rfc/rfa v2] Use ARM exception tables as GDB unwinder
- References: <201012011607.oB1G7bYA024674@d06av02.portsmouth.uk.ibm.com>
On Wed, Dec 01, 2010 at 05:07:37PM +0100, Ulrich Weigand wrote:
> - We now only trust ARM unwind info at call sites, not at arbitrary
> PC locations. (Note that we consider system calls to be call sites
> as well for this purpose, see the comment in the code.)
That's a clever compromise; I like it. We lose a bit with
asynchronous signals, but it's a very small bit since you consider
system calls as call sites too.
> @@ -1899,6 +1909,744 @@ struct frame_unwind arm_prologue_unwind
> default_frame_sniffer
> };
>
> +/* Maintain a list of ARM exception table entries per objfile, similar to the
> + list of mapping symbols. We only cache entries for standard ARM-defined
> + personality routines; the cache will contains only the frame unwinding
> + instructions associated with the entry (not the descriptors). */
"will contain" or "contains".
Otherwise, looks good to me.
--
Daniel Jacobowitz
CodeSourcery