From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from barracuda.ebox.ca (barracuda.ebox.ca [96.127.255.19]) by sourceware.org (Postfix) with ESMTPS id B5CB83858438 for ; Tue, 18 Jan 2022 04:09:48 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.1 sourceware.org B5CB83858438 X-ASG-Debug-ID: 1642478978-0c856e06ad131bf0001-fS2M51 Received: from smtp.ebox.ca (smtp.ebox.ca [96.127.255.82]) by barracuda.ebox.ca with ESMTP id HsJotmWH7b1UKDHP (version=TLSv1 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Mon, 17 Jan 2022 23:09:38 -0500 (EST) X-Barracuda-Envelope-From: simon.marchi@polymtl.ca X-Barracuda-RBL-Trusted-Forwarder: 96.127.255.82 Received: from simark.localdomain (192-222-157-6.qc.cable.ebox.net [192.222.157.6]) by smtp.ebox.ca (Postfix) with ESMTP id 6AB9D441D65; Mon, 17 Jan 2022 23:09:38 -0500 (EST) From: Simon Marchi X-Barracuda-RBL-IP: 192.222.157.6 X-Barracuda-Effective-Source-IP: 192-222-157-6.qc.cable.ebox.net[192.222.157.6] X-Barracuda-Apparent-Source-IP: 192.222.157.6 To: gdb-patches@sourceware.org Subject: [PATCH v2 2/9] gdb/linux-nat: remove check based on current_inferior in linux_handle_extended_wait Date: Mon, 17 Jan 2022 23:09:30 -0500 X-ASG-Orig-Subj: [PATCH v2 2/9] gdb/linux-nat: remove check based on current_inferior in linux_handle_extended_wait Message-Id: <20220118040937.730282-3-simon.marchi@polymtl.ca> X-Mailer: git-send-email 2.34.1 In-Reply-To: <20220118040937.730282-1-simon.marchi@polymtl.ca> References: <20220118040937.730282-1-simon.marchi@polymtl.ca> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Barracuda-Connect: smtp.ebox.ca[96.127.255.82] X-Barracuda-Start-Time: 1642478978 X-Barracuda-Encrypted: DHE-RSA-AES256-SHA X-Barracuda-URL: https://96.127.255.19:443/cgi-mod/mark.cgi X-Virus-Scanned: by bsmtpd at ebox.ca X-Barracuda-Scan-Msg-Size: 10113 X-Barracuda-BRTS-Status: 1 X-Barracuda-Spam-Score: 0.00 X-Barracuda-Spam-Status: No, SCORE=0.00 using global scores of TAG_LEVEL=1000.0 QUARANTINE_LEVEL=1000.0 KILL_LEVEL=8.0 tests= X-Barracuda-Spam-Report: Code version 3.2, rules version 3.2.3.95404 Rule breakdown below pts rule name description ---- ---------------------- -------------------------------------------------- X-Spam-Status: No, score=-15.8 required=5.0 tests=BAYES_00, GIT_PATCH_0, KAM_DMARC_QUARANTINE, KAM_DMARC_STATUS, KAM_SHORT, RCVD_IN_DNSWL_LOW, SPF_HELO_NONE, SPF_SOFTFAIL, TXREP autolearn=ham autolearn_force=no version=3.4.4 X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on server2.sourceware.org X-BeenThere: gdb-patches@sourceware.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Gdb-patches mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 18 Jan 2022 04:09:51 -0000 The check removed by this patch, using current_inferior, looks wrong. When debugging multiple inferiors with the Linux native target and linux_handle_extended_wait is called, there's no guarantee about which is the current inferior. The vfork-done event we receive could be for any inferior. If the vfork-done event is for a non-current inferior, we end up wrongfully ignoring it. As a result, the core never processes a TARGET_WAITKIND_VFORK_DONE event, program_space::breakpoints_not_allowed is never cleared, and breakpoints are never reinserted. However, because the Linux native target decided to ignore the event, it resumed the thread - while breakpoints out. And that's bad. The proposed fix is to remove this check. Always report vfork-done events and let infrun's logic decide if it should be ignored. We don't save much cycles by filtering the event here. Add a test that sets replicates the situation described above. See comments in the test for more details. Change-Id: Ibe33c1716c3602e847be6c2093120696f2286fbf --- gdb/linux-nat.c | 17 +-- .../gdb.threads/vfork-multi-inferior-sleep.c | 25 ++++ .../gdb.threads/vfork-multi-inferior.c | 55 +++++++++ .../gdb.threads/vfork-multi-inferior.exp | 115 ++++++++++++++++++ 4 files changed, 199 insertions(+), 13 deletions(-) create mode 100644 gdb/testsuite/gdb.threads/vfork-multi-inferior-sleep.c create mode 100644 gdb/testsuite/gdb.threads/vfork-multi-inferior.c create mode 100644 gdb/testsuite/gdb.threads/vfork-multi-inferior.exp diff --git a/gdb/linux-nat.c b/gdb/linux-nat.c index b926fb5eba9e..7de4da03a34f 100644 --- a/gdb/linux-nat.c +++ b/gdb/linux-nat.c @@ -2096,20 +2096,11 @@ linux_handle_extended_wait (struct lwp_info *lp, int status) if (event == PTRACE_EVENT_VFORK_DONE) { - if (current_inferior ()->waiting_for_vfork_done) - { - linux_nat_debug_printf - ("Got expected PTRACE_EVENT_VFORK_DONE from LWP %ld: stopping", - lp->ptid.lwp ()); - - ourstatus->set_vfork_done (); - return 0; - } - linux_nat_debug_printf - ("Got PTRACE_EVENT_VFORK_DONE from LWP %ld: ignoring", lp->ptid.lwp ()); - - return 1; + ("Got expected PTRACE_EVENT_VFORK_DONE from LWP %ld", + lp->ptid.lwp ()); + ourstatus->set_vfork_done (); + return 0; } internal_error (__FILE__, __LINE__, diff --git a/gdb/testsuite/gdb.threads/vfork-multi-inferior-sleep.c b/gdb/testsuite/gdb.threads/vfork-multi-inferior-sleep.c new file mode 100644 index 000000000000..a26ae07d53b7 --- /dev/null +++ b/gdb/testsuite/gdb.threads/vfork-multi-inferior-sleep.c @@ -0,0 +1,25 @@ +/* This testcase is part of GDB, the GNU debugger. + + Copyright 2022 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 . */ + +#include + +int +main (void) +{ + sleep (30); + return 0; +} diff --git a/gdb/testsuite/gdb.threads/vfork-multi-inferior.c b/gdb/testsuite/gdb.threads/vfork-multi-inferior.c new file mode 100644 index 000000000000..1b7b00d3c755 --- /dev/null +++ b/gdb/testsuite/gdb.threads/vfork-multi-inferior.c @@ -0,0 +1,55 @@ +/* This testcase is part of GDB, the GNU debugger. + + Copyright 2022 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 . */ + +#include +#include +#include + +static void +should_break_here (void) +{ +} + +int +main (void) +{ + int i; + + for (i = 0; i < NR_LOOPS; i++) + { + int pid = vfork (); + + if (pid != 0) + { + /* Parent */ + int stat; + int ret = waitpid (pid, &stat, 0); + assert (ret == pid); + assert (WIFEXITED (stat)); + assert (WEXITSTATUS (stat) == 12); + + should_break_here (); + } + else + { + /* Child */ + _exit(12); + } + } + + return 0; +} diff --git a/gdb/testsuite/gdb.threads/vfork-multi-inferior.exp b/gdb/testsuite/gdb.threads/vfork-multi-inferior.exp new file mode 100644 index 000000000000..22a1476eaacc --- /dev/null +++ b/gdb/testsuite/gdb.threads/vfork-multi-inferior.exp @@ -0,0 +1,115 @@ +# Copyright 2020-2022 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 . + +# Test handling a vfork while another inferior is running. The bug that +# prompted writing this test case was in the Linux native target. The target +# assumed that the vfork-done event it received was for the current inferior +# (an invalid assumption, the current inferior is the one randomly selected by +# do_target_wait (at the time of writing). This caused the target to drop the +# vfork-done event, because it was seen as unneeded and to restart the thread +# as if nothing happened. This however resulted in the thread running with +# breakpoints not inserted. +# +# To catch the bug, this test verifies that we can hit a breakpoint after a +# vfork call, while a second inferior runs in the background. + +if [use_gdb_stub] { + unsupported "test uses multiple inferiors" + return +} + +standard_testfile .c -sleep.c + +set srcfile_sleep $srcfile2 +set binfile_sleep ${binfile}-sleep + +# The reproducibility of the bug depends on which inferior randomly selects in +# do_target_wait when consuming the vfork-done event. Since GDB doesn't call +# srand(), we are likely to always see the same sequence of inferior selected by +# do_target_wait, which can hide the bug if you are not "lucky". To work +# around that, call vfork and hit the breakpoint in a loop, it makes it +# somewhat likely that the wrong inferior will be selected eventually. +set nr_loops 20 + +# Compile the main program that calls vfork and hits a breakpoint. +set opts [list debug additional_flags=-DNR_LOOPS=$nr_loops] +if { [gdb_compile "${srcdir}/${subdir}/${srcfile}" "${binfile}" executable \ + $opts] != "" } { + untested "failed to compile" + return -1 +} + +# Compile the secondary program, which just sleeps. +if { [gdb_compile "${srcdir}/${subdir}/${srcfile_sleep}" "${binfile_sleep}" executable \ + {debug}] != "" } { + untested "failed to compile" + return -1 +} + +# We exercise two methods of getting a second inferior to execute while the +# first one vforks. METHOD can be: +# +# - non-stop: start GDB with non-stop on and run the second inferior in +# background. +# - schedule-multiple: set "schedule-multiple on", this will make "continue" +# resume both inferiors. +proc do_test {method} { + save_vars { ::GDBFLAGS } { + if { $method == "non-stop" } { + append ::GDBFLAGS " -ex \"set non-stop on\"" + } + clean_restart + } + + # Start the second inferior in background. + gdb_test "add-inferior" "Added inferior 2.*" + gdb_test "inferior 2" "Switching to inferior 2 .*" + gdb_file_cmd ${::binfile_sleep} + if { $method == "non-stop" } { + gdb_test "run &" "Starting program: .*" "run inferior 2" + } else { + gdb_test "start" "Temporary breakpoint $::decimal, main .*" \ + "start inferior 2" + } + + # Start the first inferior. + gdb_test "inferior 1" "Switching to inferior 1 .*" + gdb_file_cmd ${::binfile} + gdb_test "break should_break_here" "Breakpoint $::decimal at .*" + gdb_test "start" "Thread 1.1 .* hit Temporary breakpoint.*" \ + "start inferior 1" + + # Only enable schedule-multiple this late, because of: + # https://sourceware.org/bugzilla/show_bug.cgi?id=28777 + if { $method == "schedule-multiple" } { + gdb_test_no_output "set schedule-multiple on" + } + + + # Continue over vfork and until the breakpoint. The number of loops here + # matches the number of loops in the program. So if a breakpoint is missed + # at some point, a "continue" will wrongfully continue until the end of the + # program, which will fail the test. + for {set i 0} {$i < $::nr_loops} {incr i} { + with_test_prefix "i=$i" { + gdb_test "continue" \ + "Thread 1.1 .* hit Breakpoint $::decimal, should_break_here.*" + } + } +} + +foreach_with_prefix method {schedule-multiple non-stop} { + do_test $method +} -- 2.34.1