From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 13833 invoked by alias); 1 Jan 2004 23:16:07 -0000 Mailing-List: contact gdb-help@sources.redhat.com; run by ezmlm Precedence: bulk List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-owner@sources.redhat.com Received: (qmail 13825 invoked from network); 1 Jan 2004 23:16:04 -0000 Received: from unknown (HELO calvin.codito.com) (203.199.140.162) by sources.redhat.com with SMTP; 1 Jan 2004 23:16:04 -0000 Received: from codito.com (numenor.codito.co.in [192.168.100.52]) by calvin.codito.com (8.12.10/) with ESMTP id i01NDW1B011708; Fri, 2 Jan 2004 04:43:35 +0530 Message-ID: <3FF4A97A.4010504@codito.com> Date: Thu, 01 Jan 2004 23:16:00 -0000 From: Ramana Radhakrishnan User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 MIME-Version: 1.0 To: Daniel Jacobowitz CC: Andrew Cagney , gdb@sources.redhat.com, ankit thukral Subject: Re: [RFC] Addition of remote tracepoints to gdb / gdbserver. References: <3FF49153.50100@codito.com> <20040101214655.GB12798@nevyn.them.org> <3FF49E8A.7020403@codito.com> <20040101225507.GA14446@nevyn.them.org> In-Reply-To: <20040101225507.GA14446@nevyn.them.org> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-SW-Source: 2004-01/txt/msg00024.txt.bz2 Daniel Jacobowitz wrote: >On Fri, Jan 02, 2004 at 03:56:18AM +0530, Ramana Radhakrishnan wrote: > > >>Now the remote side would have to understand the target floating point >>format / collect the values for >>the same and hence the actual computation would have to be target >>specific .2 options >>a. do all calculations in gdbserver >>b. collect all byte sequences and take care of it in gdb. >> >>Option b seems less expensive. >> >> > >Yes, but sometimes it's not enough. In the example I gave you'd have >to do the memory reference at trace collection time. But I don't >think it's a real problem so option b is probably fine. > > It might be cheaper to take option b .,for FPU less architectures That could be another reason for sticking to option b. Also regarding the AgentExpression code Generation problem that you mentioned in your earlier mail , the byte code generated for the following actions in the agent expressions is as follows. considering 2 globals i & j and we need to collect the value of i + j at some trace point. trace someprogram point collect i + j code generated seems to be equivalent to the following Aopconst32 &i Aopref32 Aopconst32 &j Aopref32 aopadd (where aopconst32 ) and aopadd is addition . which is consistent with the manual cheers Ramana