public inbox for gdb-cvs@sourceware.org help / color / mirror / Atom feed
From: Simon Marchi <simark@sourceware.org> To: gdb-cvs@sourceware.org Subject: [binutils-gdb] gdb/linux-nat: remove check based on current_inferior in linux_handle_extended_wait Date: Thu, 31 Mar 2022 17:08:34 +0000 (GMT) [thread overview] Message-ID: <20220331170834.EEB5E385741C@sourceware.org> (raw) https://sourceware.org/git/gitweb.cgi?p=binutils-gdb.git;h=5a0c4a06eb4f4afcf885596ee97706c1045a8476 commit 5a0c4a06eb4f4afcf885596ee97706c1045a8476 Author: Simon Marchi <simon.marchi@polymtl.ca> Date: Mon Jan 17 11:27:36 2022 -0500 gdb/linux-nat: remove check based on current_inferior in linux_handle_extended_wait 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 replicates the situation described above. See comments in the test for more details. Change-Id: Ibe33c1716c3602e847be6c2093120696f2286fbf Diff: --- gdb/linux-nat.c | 17 +-- .../gdb.threads/vfork-multi-inferior-sleep.c | 25 +++++ gdb/testsuite/gdb.threads/vfork-multi-inferior.c | 55 ++++++++++ gdb/testsuite/gdb.threads/vfork-multi-inferior.exp | 115 +++++++++++++++++++++ 4 files changed, 199 insertions(+), 13 deletions(-) diff --git a/gdb/linux-nat.c b/gdb/linux-nat.c index a9b093c00a4..dff8d07d3f7 100644 --- a/gdb/linux-nat.c +++ b/gdb/linux-nat.c @@ -1990,20 +1990,11 @@ 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 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 00000000000..a26ae07d53b --- /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 <http://www.gnu.org/licenses/>. */ + +#include <unistd.h> + +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 00000000000..7428ba00f3d --- /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 <http://www.gnu.org/licenses/>. */ + +#include <unistd.h> +#include <sys/wait.h> +#include <assert.h> + +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 00000000000..22a1476eaac --- /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 <http://www.gnu.org/licenses/>. + +# 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 +}
reply other threads:[~2022-03-31 17:08 UTC|newest] Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=20220331170834.EEB5E385741C@sourceware.org \ --to=simark@sourceware.org \ --cc=gdb-cvs@sourceware.org \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).