From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 27199 invoked by alias); 13 Aug 2010 03:31:01 -0000 Mailing-List: contact archer-help@sourceware.org; run by ezmlm Sender: Precedence: bulk List-Post: List-Help: List-Subscribe: List-Id: Received: (qmail 27184 invoked by uid 22791); 13 Aug 2010 03:31:00 -0000 X-SWARE-Spam-Status: No, hits=-6.8 required=5.0 tests=AWL,BAYES_00,RCVD_IN_DNSWL_HI,SPF_HELO_PASS,T_RP_MATCHES_RCVD X-Spam-Check-By: sourceware.org Date: Fri, 13 Aug 2010 03:31:00 -0000 From: Kevin Buettner To: archer@sourceware.org Subject: Re: gdbstub initial code, v3 Message-ID: <20100812203053.3b6ba73e@mesquite.lan> In-Reply-To: <20100813005338.9AC74400E3@magilla.sf.frob.com> References: <20100811235810.GA9783@redhat.com> <20100812011113.GA13212@redhat.com> <20100812075211.65c3e0f3@mesquite.lan> <20100813005338.9AC74400E3@magilla.sf.frob.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit X-SW-Source: 2010-q3/txt/msg00106.txt.bz2 On Thu, 12 Aug 2010 17:53:38 -0700 (PDT) Roland McGrath wrote: > > I've skimmed the code. If I'm not mistaken, support for the 's', 'G', > > and 'M' commands isn't implemented yet? (That's single-step, write > > registers, and write memory, all of which must be supported by a stub.) > > Yes, that's fine. For a starting place, having only reading support is OK. > These bits are relatively trivial to fill in later as needed, so I think > made plenty of sense for Oleg not to include them yet. It certainly makes sense for Oleg to get communication between the stub and gdb sorted out first. As another early step, it also makes sense to attempt to stop the process and determine its status. Once that's sorted though, I should think that memory write and single-step will become important to implement. The 'M' (memory write) packet may be used by GDB for inserting and removing breakpoints. (Or, alternately, Z0/z0 packets may be used. I'd rather see 'M' implemented first though.) Single-step, whether implemented as an 's' packet or 'vCont;s' packet, will be needed in order to continue from a breakpoint as well as for single-stepping. Implementation of the 'G' command (register write) is probably the least important of the three that I mentioned earlier. Kevin