From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 2378 invoked by alias); 9 Mar 2008 17:34:45 -0000 Received: (qmail 2369 invoked by uid 22791); 9 Mar 2008 17:34:44 -0000 X-Spam-Check-By: sourceware.org Received: from smtpserver1.misty.com (HELO smtpserver1.misty.com) (198.137.254.134) by sourceware.org (qpsmtpd/0.31) with ESMTP; Sun, 09 Mar 2008 17:34:13 +0000 Received: from manx.misty.com (manx.misty.com [198.137.254.20]) by smtpserver1.misty.com (8.13.8/8.13.8) with ESMTP id m29HXQe5024256; Sun, 9 Mar 2008 13:33:26 -0400 (EDT) Received: from manx.misty.com (localhost [127.0.0.1]) by manx.misty.com (8.13.8+Sun/8.13.8) with ESMTP id m29HXPI8008778; Sun, 9 Mar 2008 13:33:25 -0400 (EDT) Received: from localhost (hepler@localhost) by manx.misty.com (8.13.8+Sun/8.13.8/Submit) with ESMTP id m29HXPE5008775; Sun, 9 Mar 2008 13:33:25 -0400 (EDT) Date: Sun, 09 Mar 2008 17:34:00 -0000 From: "Edward L. Hepler" To: Keith Seitz cc: insight@sources.redhat.com Subject: Re: Adding a target? In-Reply-To: Message-ID: References: <47C79954.4000502@redhat.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed Mailing-List: contact insight-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: insight-owner@sourceware.org X-SW-Source: 2008-q1/txt/msg00083.txt.bz2 Keith, I'm still trying to move up from gdb-6.5 (and insight-6.6) to insight-6.7.1 I have gotten past the GUI problem I had thanks to your earlier suggestion. I am having a problem inserting a breakpoint... This worked fine using gdb-6.5 and insight-6.6, but seems to have broken with insight-6.6. It appears that the "to_insert_breakpoint" function is not getting called. I have "watched" the "current_target" structure and see the entries for my target (vr2000) get inserted into the structure after typing the "run" command. (The old values were for the "exec" target... new values for the "vr2000" target.) What is unusual is that the "to_insert_breakpoint" structure member is not being set to the vr2000 value. The "ignore" function is instead being inserted. Later, when the breakpoint should be being inserted, the "ignore" function is called (instead of the function defined for vr2000 (which substitutes the "break" instruction for the instruction in the target memory). This worked in earlier versions... Do you know of changes that might have caused this to break? Is there some reason why the "to_insert_breakpoint" structure item would not be overloaded with the vr2000 entry when the ops target is changed? Any thoughts would be appreciated... Thanks, Ed Hepler ---------------------------------------------------------------------------- VLSI Concepts offers synthesizable processor cores for embedded control applications... See our standard products and ask about customization... ---------------------------------------------------------------------------- Dr. Edward L. Hepler President, Adjunct Professor, VLSI Concepts, Inc. Villanova University Graduate Courses: VLSI and System ECE-8440 System Design and Modeling Architecture, Design, ECE-8445 Advanced Computer Architecture and CAD ECE-8460 VLSI Design email: hepler@vlsi-concepts.com or elh@ece.villanova.edu mobile: (484) 459-1126 www: http://www.vlsi-concepts.com Read: I Cor 8:6 On Sat, 8 Mar 2008, Edward L. Hepler wrote: > > Keith, > > That seemed to do it... Many thanks... > > By the way, where is the function "code" defined? > > Thanks again, > > Ed Hepler > > ---------------------------------------------------------------------------- > VLSI Concepts offers synthesizable processor cores for embedded control > applications... See our standard products and ask about customization... > ---------------------------------------------------------------------------- > Dr. Edward L. Hepler > President, Adjunct Professor, > VLSI Concepts, Inc. Villanova University Graduate Courses: > VLSI and System ECE-8440 System Design and Modeling > Architecture, Design, ECE-8445 Advanced Computer Architecture > and CAD ECE-8460 VLSI Design > > email: hepler@vlsi-concepts.com or elh@ece.villanova.edu > mobile: (484) 459-1126 > www: http://www.vlsi-concepts.com Read: I Cor 8:6 > > On Thu, 28 Feb 2008, Keith Seitz wrote: > >> Edward L. Hepler wrote: >> >>> I'd like to use insight to drive gdb and have included the appropriate >>> gdb modifications. The "gdb-side" appears to operate properly, but >>> I'm still having some problems with the GUI portion. I'd appreciate >>> any pointers you may be able to send that would help in adding a target >>> to the GUI... An error pops up stating that "can't read >>> "PREFS_state(gdb/load/default-portname value)" no such element in array" >> >> There are two preferences that are rather closely related, "*-port" and >> "*-portname". The former is used with serial connections and the latter >> with TCP connections. In this latter case, the TCP port NUMBER is stored in >> *-portname (I don't know why!). [YUCK!] >> >> The "default" target settings are supposed to be initialized by >> TargetSelection::_init_prefs, which is initialized when the target >> selection dialog is CONSTRUCTED. Sigh. That's messed up, since >> interface.tcl relies on these values in set_target_name. Yich. I'm >> surprised this worked at all. >> >> Alas, as a quick hack, try this: >> >> 1) add a call to _init_prefs in either TargetSelection::getname or >> TargetSelection::init_target_db. I would probably do it in init_target_db >> and remove other calls to _init_pref in this file. >> >> 2) define the preference gdb/load/default-portname (set it to any integer) >> in TargetSelection::_init_prefs (follow the example for default-port at the >> end of this method). >> >> Let me know if that fixes your problem. >> >> Keith >> >