From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 10094 invoked by alias); 1 Dec 2004 16:25:43 -0000 Mailing-List: contact gdb-patches-help@sources.redhat.com; run by ezmlm Precedence: bulk List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sources.redhat.com Received: (qmail 9321 invoked from network); 1 Dec 2004 16:25:25 -0000 Received: from unknown (HELO arwen.tausq.org) (64.81.244.109) by sourceware.org with SMTP; 1 Dec 2004 16:25:25 -0000 Received: by arwen.tausq.org (Postfix, from userid 1000) id 4E4DA43810; Wed, 1 Dec 2004 08:25:24 -0800 (PST) Date: Wed, 01 Dec 2004 16:25:00 -0000 From: Randolph Chung To: gdb-patches@sources.redhat.com Subject: Re: [patch/RFA] multiarch INSTRUCTION_NULLIFIED Message-ID: <20041201162524.GD6359@tausq.org> Reply-To: Randolph Chung Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline X-GPG: for GPG key, see http://www.tausq.org/gpg.txt User-Agent: Mutt/1.5.6+20040722i X-SW-Source: 2004-12/txt/msg00019.txt.bz2 In-Reply-To: > A point that may be worth considering is that on architectures which support > jump/branch delay slots, the branch and delay slot is actually logically > executed as single compound instruction, treated by the machine as a unified > whole. i don't think this is true for hppa. for instance you can do a hardware singlestep into the delay slot of a branch insn by setting the recovery counter appropriately. this will cause an hardware interruption which the kernel can translate to a SIGTRAP. nullification is also a more general concept on hppa which doesn't only apply to branch instructions. A lot of ALU instructions can also conditioanlly nullify the next insn, which potentially could belong to a different line of source code. gcc uses this to implement if-conversion, for example. randolph -- Randolph Chung Debian GNU/Linux Developer, hppa/ia64 ports http://www.tausq.org/