From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 14184 invoked by alias); 16 Sep 2008 00:09:53 -0000 Received: (qmail 14174 invoked by uid 22791); 16 Sep 2008 00:09:51 -0000 X-Spam-Check-By: sourceware.org Received: from smtp-outbound-1.vmware.com (HELO smtp-outbound-1.vmware.com) (65.113.40.141) by sourceware.org (qpsmtpd/0.31) with ESMTP; Tue, 16 Sep 2008 00:08:57 +0000 Received: from mailhost2.vmware.com (mailhost2.vmware.com [10.16.64.160]) by smtp-outbound-1.vmware.com (Postfix) with ESMTP id 45E0F67B4; Mon, 15 Sep 2008 17:08:53 -0700 (PDT) Received: from [10.20.92.218] (promb-2s-dhcp218.eng.vmware.com [10.20.92.218]) by mailhost2.vmware.com (Postfix) with ESMTP id 390E08E599; Mon, 15 Sep 2008 17:08:53 -0700 (PDT) Message-ID: <48CEF8AB.8080706@vmware.com> Date: Tue, 16 Sep 2008 00:09:00 -0000 From: Michael Snyder User-Agent: Thunderbird 1.5.0.12 (X11/20080411) MIME-Version: 1.0 To: Michael Snyder , "gdb-patches@sourceware.org" , teawater Subject: Re: [reverse RFA] no singlestep-over-BP in reverse References: <48CEAA05.8050006@vmware.com> <20080915184245.GA21388@caradoc.them.org> In-Reply-To: <20080915184245.GA21388@caradoc.them.org> Content-Type: multipart/mixed; boundary="------------030708070107050704050102" X-IsSubscribed: yes Mailing-List: contact gdb-patches-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sourceware.org X-SW-Source: 2008-09/txt/msg00344.txt.bz2 This is a multi-part message in MIME format. --------------030708070107050704050102 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Content-length: 969 Daniel Jacobowitz wrote: > On Mon, Sep 15, 2008 at 11:31:33AM -0700, Michael Snyder wrote: >> When we're stopped at a breakpoint and we want to >> continue in reverse, we're not actually going to >> execute the instruction at the breakpoint -- we're >> going to de-execute the previous instruction. >> >> Therefore there's no need to singlestep before >> inserting breakpoints. In fact it would be a bad >> idea to do so, because if there is a breakpoint at >> the previous instruction, we WANT to hit it. >> >> Note that this patch is to be applied to the reverse branch. > > If there is a breakpoint on the previous instruction, will you hit it > before or after de-executing that instruction? It seems like this > logic should be somehow still necessary... but I can't put my finger > on when. The attached test does just that -- and it passes on three targets that can go in reverse (teawater's, gdb-freeplay, and the as yet unreleased VMware implementation). --------------030708070107050704050102 Content-Type: text/plain; name="consecutive-reverse.exp" Content-Transfer-Encoding: 7bit Content-Disposition: inline; filename="consecutive-reverse.exp" Content-length: 3284 # Copyright 2001, 2007, 2008 Free Software Foundation, Inc. # This program is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by # the Free Software Foundation; either version 3 of the License, or # (at your option) any later version. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # # You should have received a copy of the GNU General Public License # along with this program. If not, see . # Please email any bugs, comments, and/or additions to this file to: # bug-gdb@prep.ai.mit.edu # Based on a file written by Michael Snyder. (msnyder@redhat.com) # # Test breakpoints at consecutive instruction addresses. # set testfile "consecutive-reverse" set srcfile ${testfile}.c set binfile ${objdir}/${subdir}/${testfile} if { [prepare_for_testing ${testfile}.exp $testfile $srcfile] } { untested "${testfile}.exp" return -1 } runto main # FIXME 'record' command should say something so we know it's working. gdb_test "record" "" "start recording" gdb_breakpoint foo gdb_test "continue" "Breakpoint $decimal, foo .*" \ "continue to breakpoint in foo" set foo1_addr 0 set foo2_addr 0 set stop_addr 0 send_gdb "x /2i \$pc\n" gdb_expect { global hex global foo1_addr global foo2_addr global gdb_prompt -re "($hex).*\[\r\n\]+($hex).*$gdb_prompt $" { set foo1_addr $expect_out(1,string) set foo2_addr $expect_out(2,string) pass "get breakpoint address for foo" } -re ".*$gdb_prompt $" { fail "get breakpoint address for foo" return 0; } timeout { fail "get breakpoint address for foo (timeout)" return 0; } } gdb_test "break \*$foo2_addr" "Breakpoint $decimal at $foo2_addr: file .*" \ "set bp, 2nd instr" send_gdb "step\n" gdb_expect { -re "Breakpoint $decimal, ($hex) in foo.*$gdb_prompt $" { set stop_addr $expect_out(1,string) if [eval expr "$foo2_addr == $stop_addr"] then { pass "stopped at bp, 2nd instr" } else { fail "stopped at bp, 2nd instr (wrong address)" } } -re ".*$gdb_prompt $" { fail "stopped at bp, 2nd instr" } timeout { fail "stopped at bp, 2nd instr (timeout)" } } ### ### ### # Set reverse execution direction # FIXME: command needs to acknowledge, so we can test if it succeeded. gdb_test "set exec-dir reverse" "" "set reverse execution" # Now step backward and hope to hit the first breakpoint. set test_msg "stopped at bp in reverse, 1st instr" gdb_test_multiple "step" "$test_msg" { -re "Breakpoint $decimal, ($hex) in foo.*$gdb_prompt $" { set stop_addr $expect_out(1,string) if [eval expr "$foo1_addr == $stop_addr"] then { pass "$test_msg" } else { fail "$test_msg (wrong address)" } } -re "Breakpoint $decimal, foo.*$gdb_prompt $" { send_gdb "print \$pc == $foo1_addr\n" gdb_expect { -re "$decimal = 1\[\r\n\]+$gdb_prompt $" { pass "$test_msg" } -re "$decimal = 0\[\r\n\]+$gdb_prompt $" { fail "$test_msg (wrong address)" } } } -re ".*$gdb_prompt $" { fail "$test_msg" } } --------------030708070107050704050102--