From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 28796 invoked by alias); 17 Sep 2002 15:52:55 -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 28780 invoked from network); 17 Sep 2002 15:52:52 -0000 Received: from unknown (HELO twinlark.arctic.org) (208.44.199.239) by sources.redhat.com with SMTP; 17 Sep 2002 15:52:52 -0000 Received: (qmail 30201 invoked by uid 539); 17 Sep 2002 15:52:51 -0000 Received: from localhost (sendmail-bs@127.0.0.1) by localhost with SMTP; 17 Sep 2002 15:52:51 -0000 Date: Tue, 17 Sep 2002 08:52:00 -0000 From: Robert Bedichek To: Daniel Jacobowitz cc: gdb@sources.redhat.com Subject: Re: Question about TCP/IP stubs, gdb-5.2.1 In-Reply-To: <20020917153221.GA30963@nevyn.them.org> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-SW-Source: 2002-09/txt/msg00221.txt.bz2 Thanks for the pointer to gdbserver, that might be the best starting point. What I want this for a connection to a full system simulator. I used to compile simulators directly with gdb, but this time I want them separated. There are seveal reasons, but one is that I want to be able to have thousands of simulators running for days in an LSF-managed farm. I want to be able to connect to a simulator, do some debugging, and disconnect, letting the simulator run some more. This is like attaching to processes with regular gdb. So what I plan to do at this point, unless someone has an even better suggestion, is to start with gdbserver, rip out the part that talks to a Linux process, and make it so that it can be compiled with the simulator and access the simulator data structures directly. Thanks again! On Tue, 17 Sep 2002, Daniel Jacobowitz wrote: > On Tue, Sep 17, 2002 at 08:28:09AM -0700, Robert Bedichek wrote: > > > > The only TCP/IP stub I found was for Wince. I need one for Linux. Does > > anyone have a pointer to a stub, the thing that runs on the target, or the > > machine being debugged, that is closer to what I need than the Wince stub > > included in 5.2.1? (The architecture is x86) > > Are you talking about application level debugging? If so, you want > gdbserver, which is included in the GDB release. If you're talking > about kernel level debugging, then TCP/IP is generally considered > impractical from an interrupt context. > > -- > Daniel Jacobowitz > MontaVista Software Debian GNU/Linux Developer >