From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 12027 invoked by alias); 8 May 2007 22:49:27 -0000 Received: (qmail 12018 invoked by uid 22791); 8 May 2007 22:49:26 -0000 X-Spam-Check-By: sourceware.org Received: from palrel11.hp.com (HELO palrel11.hp.com) (156.153.255.246) by sourceware.org (qpsmtpd/0.31) with ESMTP; Tue, 08 May 2007 22:49:24 +0000 Received: from smtp2.cup.hp.com (smtp.cup.hp.com [15.4.32.118]) by palrel11.hp.com (Postfix) with ESMTP id 502DA36058 for ; Tue, 8 May 2007 15:49:22 -0700 (PDT) Received: from hpadl693.cup.hp.com (hpadl693.cup.hp.com [15.244.96.238]) by smtp2.cup.hp.com (Postfix) with ESMTP id 1FD4730E43D for ; Tue, 8 May 2007 22:41:19 +0000 (UTC) Received: (from cdb@localhost) by hpadl693.cup.hp.com (8.9.3 (PHNE_29774)/8.7.3 TIS Messaging 5.0) id PAA06205 for gdb@sourceware.org; Tue, 8 May 2007 15:49:22 -0700 (PDT) Date: Tue, 08 May 2007 22:49:00 -0000 From: Carl Burch Message-Id: <200705082249.PAA06205@hpadl693.cup.hp.com> To: gdb@sourceware.org Subject: Re: DWARF-2 expression error in location list entries Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Content-MD5: 16IlTSwmuclWNTdDYQs6sQ== Mailing-List: contact gdb-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-owner@sourceware.org X-SW-Source: 2007-05/txt/msg00033.txt.bz2 Daniel- > I believe the language was recently discussed on the dwarf workgroup > list. It's unclear, but the error is correct. DW_OP_regx describes > a location ("the value is in this register"), not a value ("the value is > the contents of this register"). Therefore the value may not be > stack-sized, and is not pushed on to the expression stack, so there's > nothing for you to DW_OP_deref later. > > > The expression we have in mind to add is to describe Fortran reference > > parameters in location list entries like : > > > > { 0x4022720--0x4022724: DW_OP_regx 32 DW_OP_deref } > > Isn't that just "DW_OP_bregx 32 0"? Yes, I think so. Thanks for the clear explanation. Any idea why the same situation in decode_locdesc() of dwarf2read.c is only a complaint? - Carl Burch HP WDB Team