Re: [patch v8 15/24] frame, backtrace: allow targets to supply a frame unwinder

[Pedro Alves <palves at redhat dot com>]

On 12/12/2013 09:15 AM, Markus Metzger wrote:
> Allow targets to supply their own target-specific frame unwinders; one for
> normal frames and one for tailcall frames.  If a target-specific unwinder
> is supplied, it will be chosen before any other unwinder.

These ...

> 
> diff --git a/gdb/dwarf2-frame.c b/gdb/dwarf2-frame.c
> index c4f8771..4efa7d8 100644
> --- a/gdb/dwarf2-frame.c
> +++ b/gdb/dwarf2-frame.c
> @@ -1513,16 +1513,16 @@ dwarf2_frame_cfa (struct frame_info *this_frame)
>  {
>    while (get_frame_type (this_frame) == INLINE_FRAME)
>      this_frame = get_prev_frame (this_frame);
> +  if (get_frame_unwind_stop_reason (this_frame) == UNWIND_UNAVAILABLE)
> +    throw_error (NOT_AVAILABLE_ERROR,
> +                _("can't compute CFA for this frame: "
> +                  "required registers or memory are unavailable"));
>    /* This restriction could be lifted if other unwinders are known to
>       compute the frame base in a way compatible with the DWARF
>       unwinder.  */
>    if (!frame_unwinder_is (this_frame, &dwarf2_frame_unwind)
>        && !frame_unwinder_is (this_frame, &dwarf2_tailcall_frame_unwind))
>      error (_("can't compute CFA for this frame"));
> -  if (get_frame_unwind_stop_reason (this_frame) == UNWIND_UNAVAILABLE)
> -    throw_error (NOT_AVAILABLE_ERROR,
> -		 _("can't compute CFA for this frame: "
> -		   "required registers or memory are unavailable"));
>    return get_frame_base (this_frame);
>  }

> diff --git a/gdb/frame.c b/gdb/frame.c
> index 37d780e..3f2aabb 100644
> --- a/gdb/frame.c
> +++ b/gdb/frame.c
> @@ -2507,13 +2507,10 @@ get_frame_sp (struct frame_info *this_frame)
>  enum unwind_stop_reason
>  get_frame_unwind_stop_reason (struct frame_info *frame)
>  {
> -  /* If we haven't tried to unwind past this point yet, then assume
> -     that unwinding would succeed.  */
> -  if (frame->prev_p == 0)
> -    return UNWIND_NO_REASON;
> +  /* Fill-in STOP_REASON.  */
> +  get_prev_frame_1 (frame);
> +  gdb_assert (frame->prev_p);
>  
> -  /* Otherwise, we set a reason when we succeeded (or failed) to
> -     unwind.  */
>    return frame->stop_reason;
>  }

... parts look a bit unrelated.  It'd be good to have
the rationale for this spelled out.  I can understand
why they might be necessary, but I don't see why
target vs non-unwinders unwinders would be special here?

-- 
Pedro Alves