This is the mail archive of the
mailing list for the GDB project.
Re: [PATCH 00/26] Regset rework
- From: "Ulrich Weigand" <uweigand at de dot ibm dot com>
- To: arnez at linux dot vnet dot ibm dot com (Andreas Arnez)
- Cc: gdb-patches at sourceware dot org
- Date: Mon, 22 Sep 2014 19:15:21 +0200 (CEST)
- Subject: Re: [PATCH 00/26] Regset rework
- Authentication-results: sourceware.org; auth=none
Andreas Arnez wrote:
> Now this patch series attempts the next major refactoring step in this
> area. It has two main goals:
> * Unify GDB's capabilities with respect to core files across the
> various architectures. In particular enable more targets for
> multi-arch capable core file generation support.
> * Streamline the regset support: reduce the gdbarch interface and
> reduce overall complexity.
> One important means to achieve these goals is by merging two existing
> gdbarch interfaces into one. Currently, the gdbarch variable
> 'core_regset_sections' enumerates the core file register notes to be
> handled by GDB; it specifies each note section's BFD name, size, and
> human-friendly name. In addition, the gdbarch method
> 'regset_from_core_section' translates a BFD name to a register set
> definition. Both of these interfaces are merged into a single gdbarch
> method 'iterate_over_regset_sections', which enumerates the core file
> notes *and* provides the register set definition for each.
I've reviewed the patch set, and I agree with the general direction,
and the implementation looks good to me as well.
It would still be good to get some platform maintainer feedback (and
testing), so I'd like to wait another week or so before approving
Dr. Ulrich Weigand
GNU/Linux compilers and toolchain