From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 18095 invoked by alias); 10 May 2009 18:33:31 -0000 Received: (qmail 18086 invoked by uid 22791); 10 May 2009 18:33:30 -0000 X-SWARE-Spam-Status: No, hits=-1.5 required=5.0 tests=AWL,BAYES_00,J_CHICKENPOX_34,J_CHICKENPOX_36,J_CHICKENPOX_37,SPF_HELO_PASS,SPF_PASS X-Spam-Check-By: sourceware.org Received: from mx2.redhat.com (HELO mx2.redhat.com) (66.187.237.31) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Sun, 10 May 2009 18:33:23 +0000 Received: from int-mx2.corp.redhat.com (int-mx2.corp.redhat.com [172.16.27.26]) by mx2.redhat.com (8.13.8/8.13.8) with ESMTP id n4AIWxbw022613; Sun, 10 May 2009 14:32:59 -0400 Received: from ns3.rdu.redhat.com (ns3.rdu.redhat.com [10.11.255.199]) by int-mx2.corp.redhat.com (8.13.1/8.13.1) with ESMTP id n4AIWwkx030210; Sun, 10 May 2009 14:32:59 -0400 Received: from host0.dyn.jankratochvil.net (sebastian-int.corp.redhat.com [172.16.52.221]) by ns3.rdu.redhat.com (8.13.8/8.13.8) with ESMTP id n4AIWu4P005286; Sun, 10 May 2009 14:32:58 -0400 Received: from host0.dyn.jankratochvil.net (localhost [127.0.0.1]) by host0.dyn.jankratochvil.net (8.14.3/8.14.3) with ESMTP id n4AIWsVg018643; Sun, 10 May 2009 20:32:55 +0200 Received: (from jkratoch@localhost) by host0.dyn.jankratochvil.net (8.14.3/8.14.3/Submit) id n4AIWnQF018629; Sun, 10 May 2009 20:32:49 +0200 Date: Sun, 10 May 2009 18:33:00 -0000 From: Jan Kratochvil To: Joel Brobecker Cc: Tom Tromey , gdb-patches@sourceware.org Subject: Re: [patch] Fix internal error on breaking at a multi-locations caller Message-ID: <20090510183249.GA12441@host0.dyn.jankratochvil.net> References: <20090309220736.GA27259@host0.dyn.jankratochvil.net> <20090428203235.GG31821@adacore.com> <20090501091942.GA8465@host0.dyn.jankratochvil.net> <20090501154759.GF10734@adacore.com> <20090501173212.GA31331@host0.dyn.jankratochvil.net> <20090506165143.GM10734@adacore.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20090506165143.GM10734@adacore.com> User-Agent: Mutt/1.5.19 (2009-01-05) 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: 2009-05/txt/msg00196.txt.bz2 Hi Joel, still requesting approval now just for the ChangeLog entry. On Wed, 06 May 2009 18:51:43 +0200, Joel Brobecker wrote: > The code part is OK, but I was a little confused by the comment. I agree my comment was misleading, used now yours. > > +# PC should not be at the boundary of source lines to make the original bug > > +# exploitable. > > + > > +set test "p/x \$pc" > > +set pc {} > > +gdb_test_multiple $test $test { > > + -re "\\$\[0-9\]+ = (0x\[0-9a-f\]+)\r\n$gdb_prompt $" { > > + set pc $expect_out(1,string) > > + pass $test > > + } > > +} > > + > > +set test "info line" > > +set end {} > > +gdb_test_multiple $test $test { > > + -re "Line \[0-9\]+ of .* starts at address 0x\[0-9a-f\]+.* and ends at (0x\[0-9a-f\]+).*\\.\r\n$gdb_prompt $" { > > + set end $expect_out(1,string) > > + pass $test > > + } > > +} > > + > > +set test "caller line has trailing code" > > +if {$pc != $end} { > > + pass $test > > +} else { > > + fail $test > > +} > > I don't think this is right - the fail here is conditional on the code > generated by the compiler on the given platform. If the return address > points at the beginning of the next line of code, then our testcase > won't allow us to test the issue. But there's no real failure > demonstrated at this point. I agree just chose it as the least worse way out: FAIL did mean a bug in the testcase itself as it did not force the system compiler to produce code acceptable for the testing. Some UNRESOLVED/XFAIL would be probably the right result code but one may not give enough priority for such regression. > My feeling on this is that we should just do the testing you do after > without worrying whether the break address is at the beginning of > a line or not. So I'd just ditch the above, and keep the rest of > the testcase as is. OK, removed this part of code as it should never happen. Put there now: +# PC should not be now at the boundary of source lines to make the original bug +# exploitable. The GLOBAL_X variable exists in the source for this purpose. Thanks, Jan gdb/ 2009-05-10 Joel Brobecker Fix internal error on breaking at a multi-locations caller source line. * breakpoint.c (parse_breakpoint_sals): Set EXPLICIT_PC for the `break' command with no parameters. gdb/testsuite/ 2009-05-10 Jan Kratochvil * gdb.cp/expand-sals.exp, gdb.cp/expand-sals.cc: New. --- gdb/breakpoint.c 5 May 2009 13:24:48 -0000 1.393 +++ gdb/breakpoint.c 10 May 2009 18:22:47 -0000 @@ -5494,6 +5494,14 @@ parse_breakpoint_sals (char **address, sal.line = default_breakpoint_line; sal.symtab = default_breakpoint_symtab; sal.section = find_pc_overlay (sal.pc); + + /* "break" without arguments is equivalent to "break *PC" where PC is + the default_breakpoint_address. So make sure to set + sal.explicit_pc to prevent GDB from trying to expand the list of + sals to include all other instances with the same symtab and line. + */ + sal.explicit_pc = 1; + sals->sals[0] = sal; sals->nelts = 1; } --- /dev/null 1 Jan 1970 00:00:00 -0000 +++ gdb/testsuite/gdb.cp/expand-sals.cc 10 May 2009 18:22:47 -0000 @@ -0,0 +1,53 @@ +/* This testcase is part of GDB, the GNU debugger. + + Copyright (C) 2009 Free Software Foundation, Inc. + + This file is part of GDB. + + 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 . */ + +int +func () +{ + return 42; /* func-line */ +} + +volatile int global_x; + +class A +{ +public: + A () + { + global_x = func (); /* caller-line */ + } +}; + +/* class B is here just to make the `func' calling line above having multiple + instances - multiple locations. Template cannot be used as its instances + would have different function names which get discarded by GDB + expand_line_sal_maybe. */ + +class B : public A +{ +}; + +int +main (void) +{ + A a; + B b; + + return 0; /* exit-line */ +} --- /dev/null 1 Jan 1970 00:00:00 -0000 +++ gdb/testsuite/gdb.cp/expand-sals.exp 10 May 2009 18:22:47 -0000 @@ -0,0 +1,58 @@ +# Copyright 2009 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 . + +if { [skip_cplus_tests] } { continue } + +set srcfile expand-sals.cc +if { [prepare_for_testing expand-sals.exp expand-sals $srcfile {debug c++}] } { + return -1 +} +if ![runto_main] { + return -1 +} + +gdb_breakpoint [gdb_get_line_number "exit-line"] + +gdb_breakpoint [gdb_get_line_number "func-line"] +gdb_continue_to_breakpoint "func" ".*func-line.*" + +gdb_test "up" "caller-line.*" + +# PC should not be now at the boundary of source lines to make the original bug +# exploitable. The GLOBAL_X variable exists in the source for this purpose. + +# Original problem was an internal error here. +set test "break" +gdb_test_multiple $test $test { + -re "Breakpoint \[0-9\]+ at .*, line \[0-9\]+\\. \\(\[2-9\] locations\\)\r\n$gdb_prompt $" { + fail $test + } + -re "Breakpoint \[0-9\]+ at .*, line \[0-9\]+\\.\r\n$gdb_prompt $" { + pass $test + } +} + +gdb_continue_to_breakpoint "caller" ".*caller-line.*" + +# Test GDB caught this return call and not the next one through B::B() +gdb_test "bt" \ + "#0 \[^\r\n\]* A \[^\r\n\]*\r\n#1 \[^\r\n\]* main \[^\r\n\]*" \ + "bt from A" + +gdb_continue_to_breakpoint "next caller func" ".*func-line.*" + +# Verify GDB really could not catch any other breakpoint location. + +gdb_continue_to_breakpoint "uncaught return" ".*exit-line.*"