From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <ac131313@redhat.com>
Mailing-List: contact gdb-patches-help@sources.redhat.com; run by ezmlm
Received: (qmail 3221 invoked from network); 10 Jan 2003 20:10:46 -0000
Received: from unknown (HELO localhost.redhat.com) (216.138.202.10)
  by 209.249.29.67 with SMTP; 10 Jan 2003 20:10:46 -0000
Received: from redhat.com (localhost [127.0.0.1])
	by localhost.redhat.com (Postfix) with ESMTP
	id BA7FC3E02; Fri, 10 Jan 2003 15:10:35 -0500 (EST)
Message-ID: <3E1F28BB.9080704@redhat.com>
Date: Fri, 10 Jan 2003 20:10:00 -0000
From: Andrew Cagney <ac131313@redhat.com>
User-Agent: Mozilla/5.0 (X11; U; NetBSD macppc; en-US; rv:1.0.1) Gecko/20021211
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Daniel Jacobowitz <drow@mvista.com>
Cc: gdb-patches@sources.redhat.com
Subject: Re: RFC: Mostly kill FRAME_CHAIN_VALID, add user knob
References: <20021226191541.GA8483@nevyn.them.org>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-SW-Source: 2003-01/txt/msg00407.txt.bz2

FYI, on the unwind branch, I'm adding the below to get_prev_frame()

   /* There is always a frame.  If this assertion fails, suspect that
      something should be calling get_selected_frame() or
      get_current_frame().  */
   gdb_assert (next_frame != NULL);

   if (next_frame->level >= 0
       /* && !backtrace_below_main */
       && inside_main_func (next_frame->pc))
     /* Don't unwind past main(), always unwind the sentinel frame.
        Note, this is done _before_ the frame has been marked as
        previously unwound.  That way if the user later decides to
        allow unwinds past main(), it can just happen.  */
     return 0;

   /* Only try to do the unwind once.  */
   if (next_frame->prev_p)
     return next_frame->prev;
   next_frame->prev_p = 1;

Should eliminate that need to flush the frame cache everytime that 
command is issued (should be able to merge it into current code as well).

Looking at the other checks:

   /* If we're already inside the entry function for the main objfile, 
then it isn't valid.  */
   if (inside_entry_func (get_frame_pc (fi)))
     return 0;

I'd better add that one as well (but, I think, after ->prev_p).   Note 
that this and the test below do the same thing, so only one is needed. 
It is just that one stops things a frame later.

   /* If we're inside the entry file, it isn't valid.  */
   /* NOTE/drow 2002-12-25: should there be a way to disable this check? 
  it assumes a single small entry file, and the way some debug readers 
(e.g. dbxread) figure out which object is the entry file is somewhat 
hokey.  */
   if (inside_entry_file (frame_pc_unwind (fi)))
     return 0;

Should this one be dropped?  If the user specified unwind past main, 
then they problably want _start() included in the backtrace.  It's 
presence also makes the other inside_entry_file() test redundant.

Andrew

PS: The command/variable should be `backtrace-*above*-main', or perhaphs 
`backtrace-before-main' (before ~= prev).