From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 26245 invoked by alias); 18 Apr 2019 09:12:07 -0000 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 Received: (qmail 26235 invoked by uid 89); 18 Apr 2019 09:12:06 -0000 Authentication-Results: sourceware.org; auth=none X-Spam-SWARE-Status: No, score=-26.9 required=5.0 tests=BAYES_00,GIT_PATCH_0,GIT_PATCH_1,GIT_PATCH_2,GIT_PATCH_3,KAM_SHORT,RCVD_IN_DNSWL_NONE autolearn=ham version=3.3.1 spammy= X-HELO: mail-wr1-f65.google.com Received: from mail-wr1-f65.google.com (HELO mail-wr1-f65.google.com) (209.85.221.65) by sourceware.org (qpsmtpd/0.93/v0.84-503-g423c35a) with ESMTP; Thu, 18 Apr 2019 09:12:04 +0000 Received: by mail-wr1-f65.google.com with SMTP id o12so2043820wrn.2 for ; Thu, 18 Apr 2019 02:12:03 -0700 (PDT) Return-Path: Received: from ?IPv6:2001:8a0:f913:f700:4c97:6d52:2cea:997b? ([2001:8a0:f913:f700:4c97:6d52:2cea:997b]) by smtp.gmail.com with ESMTPSA id i2sm1884247wmg.47.2019.04.18.02.12.00 (version=TLS1_3 cipher=AEAD-AES128-GCM-SHA256 bits=128/128); Thu, 18 Apr 2019 02:12:01 -0700 (PDT) Subject: Re: [PATCH] Handle vfork in thread with follow-fork-mode child To: Tom de Vries , gdb-patches@sourceware.org References: <20190416150652.GA4805@delia> From: Pedro Alves Message-ID: <3bb18786-b84d-ebdb-233a-92237f3c7764@redhat.com> Date: Thu, 18 Apr 2019 09:12:00 -0000 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101 Thunderbird/60.2.1 MIME-Version: 1.0 In-Reply-To: Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit X-SW-Source: 2019-04/txt/msg00311.txt.bz2 On 4/18/19 9:02 AM, Tom de Vries wrote: > On 17-04-19 19:45, Pedro Alves wrote: > >> + inf->vfork_parent = NULL; > > I was checking this and found dereferences of inf->vfork_parent after it > was set to NULL here: > ... > else if (exec) > { > ... > resume_parent = inf->vfork_parent->pid; > ... > and here: > ... > else > { > ... > clone_program_space (pspace, inf->vfork_parent->pspace); > ... > > To confirm, I did another testrun with aborts at the start of the > blocks, and I found no regressions. So, either this is dead code, or we > need test-cases that trigger these paths. Indeed, I noticed this yesterday, fixed it & wrote the extra testing, but then had to leave for the day. Here is the updated patch with extra testing, which involved moving the body of the testcases to a procedure, and then testing with both "set detach-on-fork" "on" and "off". WDYT? >From 575fecd185d07cd0d2f9d9aed5325e7b09b675e0 Mon Sep 17 00:00:00 2001 From: Pedro Alves Date: Thu, 18 Apr 2019 09:57:45 +0100 Subject: [PATCH] [gdb] Handle vfork in thread with follow-fork-mode child When debugging any of the testcases added by this commit, which do a vfork in a thread with "set follow-fork-mode child" + "set detach-on-fork on", we run into this assertion: ... src/gdb/nat/x86-linux-dregs.c:146: internal-error: \ void x86_linux_update_debug_registers(lwp_info*): \ Assertion `lwp_is_stopped (lwp)' failed. ... The assert is caused by the following: the vfork-child exit or exec event is handled by handle_vfork_child_exec_or_exit, which calls target_detach to detach from the vfork parent. During target_detach we call linux_nat_target::detach, which: #1 - stops all the threads #2 - waits for all the threads to be stopped #3 - detaches all the threads However, during the second step we run into this code in stop_wait_callback: ... /* If this is a vfork parent, bail out, it is not going to report any SIGSTOP until the vfork is done with. */ if (inf->vfork_child != NULL) return 0; ... and we don't wait for the threads to be stopped, which results in this assert in x86_linux_update_debug_registers triggering during the third step: ... gdb_assert (lwp_is_stopped (lwp)); ... The fix is to reset the vfork parent's vfork_child field before calling target_detach in handle_vfork_child_exec_or_exit. There's already similar code for the other paths handled by handle_vfork_child_exec_or_exit, so this commit refactors the code a bit so that all paths share the same code. The new tests cover both a vfork child exiting, and a vfork child execing, since both cases would trigger the assertion. The new testcases also exercise following the vfork children with "set detach-on-fork off", since it doesn't seem to be tested anywhere. Tested on x86_64-linux, using native and native-gdbserver. gdb/ChangeLog: yyyy-mm-dd Tom de Vries Pedro Alves PR gdb/24454 * infrun.c (handle_vfork_child_exec_or_exit): Reset vfork parent's vfork_child field before calling target_detach. gdb/testsuite/ChangeLog: yyyy-mm-dd Tom de Vries Pedro Alves PR gdb/24454 * gdb.threads/vfork-follow-child-exec.c: New file. * gdb.threads/vfork-follow-child-exec.exp: New file. * gdb.threads/vfork-follow-child-exit.c: New file. * gdb.threads/vfork-follow-child-exit.exp: New file. --- gdb/infrun.c | 31 +++++----- .../gdb.threads/vfork-follow-child-exec.c | 66 ++++++++++++++++++++++ .../gdb.threads/vfork-follow-child-exec.exp | 64 +++++++++++++++++++++ .../gdb.threads/vfork-follow-child-exit.c | 52 +++++++++++++++++ .../gdb.threads/vfork-follow-child-exit.exp | 60 ++++++++++++++++++++ 5 files changed, 256 insertions(+), 17 deletions(-) create mode 100644 gdb/testsuite/gdb.threads/vfork-follow-child-exec.c create mode 100644 gdb/testsuite/gdb.threads/vfork-follow-child-exec.exp create mode 100644 gdb/testsuite/gdb.threads/vfork-follow-child-exit.c create mode 100644 gdb/testsuite/gdb.threads/vfork-follow-child-exit.exp diff --git a/gdb/infrun.c b/gdb/infrun.c index 37713b24fee..37df561de0b 100644 --- a/gdb/infrun.c +++ b/gdb/infrun.c @@ -915,10 +915,14 @@ handle_vfork_child_exec_or_exit (int exec) int resume_parent = -1; /* This exec or exit marks the end of the shared memory region - between the parent and the child. If the user wanted to - detach from the parent, now is the time. */ + between the parent and the child. Break the bonds. */ + inferior *vfork_parent = inf->vfork_parent; + inf->vfork_parent->vfork_child = NULL; + inf->vfork_parent = NULL; - if (inf->vfork_parent->pending_detach) + /* If the user wanted to detach from the parent, now is the + time. */ + if (vfork_parent->pending_detach) { struct thread_info *tp; struct program_space *pspace; @@ -926,7 +930,7 @@ handle_vfork_child_exec_or_exit (int exec) /* follow-fork child, detach-on-fork on. */ - inf->vfork_parent->pending_detach = 0; + vfork_parent->pending_detach = 0; gdb::optional maybe_restore_inferior; @@ -941,7 +945,7 @@ handle_vfork_child_exec_or_exit (int exec) maybe_restore_thread.emplace (); /* We're letting loose of the parent. */ - tp = any_live_thread_of_inferior (inf->vfork_parent); + tp = any_live_thread_of_inferior (vfork_parent); switch_to_thread (tp); /* We're about to detach from the parent, which implicitly @@ -964,7 +968,7 @@ handle_vfork_child_exec_or_exit (int exec) if (print_inferior_events) { std::string pidstr - = target_pid_to_str (ptid_t (inf->vfork_parent->pid)); + = target_pid_to_str (ptid_t (vfork_parent->pid)); target_terminal::ours_for_output (); @@ -982,7 +986,7 @@ handle_vfork_child_exec_or_exit (int exec) } } - target_detach (inf->vfork_parent, 0); + target_detach (vfork_parent, 0); /* Put it back. */ inf->pspace = pspace; @@ -997,10 +1001,7 @@ handle_vfork_child_exec_or_exit (int exec) inf->removable = 1; set_current_program_space (inf->pspace); - resume_parent = inf->vfork_parent->pid; - - /* Break the bonds. */ - inf->vfork_parent->vfork_child = NULL; + resume_parent = vfork_parent->pid; } else { @@ -1030,17 +1031,13 @@ handle_vfork_child_exec_or_exit (int exec) set_current_program_space (pspace); inf->removable = 1; inf->symfile_flags = SYMFILE_NO_READ; - clone_program_space (pspace, inf->vfork_parent->pspace); + clone_program_space (pspace, vfork_parent->pspace); inf->pspace = pspace; inf->aspace = pspace->aspace; - resume_parent = inf->vfork_parent->pid; - /* Break the bonds. */ - inf->vfork_parent->vfork_child = NULL; + resume_parent = vfork_parent->pid; } - inf->vfork_parent = NULL; - gdb_assert (current_program_space == inf->pspace); if (non_stop && resume_parent != -1) diff --git a/gdb/testsuite/gdb.threads/vfork-follow-child-exec.c b/gdb/testsuite/gdb.threads/vfork-follow-child-exec.c new file mode 100644 index 00000000000..80632d1772f --- /dev/null +++ b/gdb/testsuite/gdb.threads/vfork-follow-child-exec.c @@ -0,0 +1,66 @@ +/* This testcase is part of GDB, the GNU debugger. + + Copyright 2019 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 +#include +#include + +static char *program_name; + +static void * +f (void *arg) +{ + int res = vfork (); + + if (res == -1) + { + perror ("vfork"); + return NULL; + } + else if (res == 0) + { + /* Child. */ + execl (program_name, program_name, "1", NULL); + perror ("exec"); + abort (); + } + else + { + /* Parent. */ + return NULL; + } +} + +int +main (int argc, char **argv) +{ + pthread_t tid; + + if (argc > 1) + { + /* Getting here via execl. */ + return 0; + } + + program_name = argv[0]; + + pthread_create (&tid, NULL, f, NULL); + pthread_join (tid, NULL); + return 0; +} diff --git a/gdb/testsuite/gdb.threads/vfork-follow-child-exec.exp b/gdb/testsuite/gdb.threads/vfork-follow-child-exec.exp new file mode 100644 index 00000000000..5a28715fa0d --- /dev/null +++ b/gdb/testsuite/gdb.threads/vfork-follow-child-exec.exp @@ -0,0 +1,64 @@ +# Copyright (C) 2019 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 following a vfork child that execs, when the vfork parent is a +# threaded program, and it's a non-main thread that vforks. + +standard_testfile + +if {[build_executable "failed to prepare" $testfile $srcfile {debug pthreads}]} { + return -1 +} + +# DETACH indicates whether "set detach-on-fork" is enabled. It is +# either "on" or "off". + +proc test_vfork {detach} { + global binfile + + clean_restart $binfile + + if ![runto_main] then { + fail "can't run to main" + return 0 + } + + delete_breakpoints + + gdb_test_no_output "set follow-fork-mode child" + gdb_test_no_output "set detach-on-fork $detach" + + if {$detach == "off"} { + gdb_test "continue" \ + [multi_line \ + "Attaching after .* vfork to child .*" \ + ".*New inferior 2 .*" \ + ".* is executing new program: .*" \ + ".*Inferior 2 .* exited normally.*"] + } else { + gdb_test "continue" \ + [multi_line \ + "Attaching after .* vfork to child .*" \ + ".*New inferior 2 .*" \ + ".*Detaching vfork parent process .* after child exec.*" \ + ".*Inferior 1 .* detached.*" \ + ".*is executing new program: .*" \ + ".*Inferior 2 .*exited normally.*"] + } +} + +foreach_with_prefix detach-on-fork {"off" "on"} { + test_vfork ${detach-on-fork} +} diff --git a/gdb/testsuite/gdb.threads/vfork-follow-child-exit.c b/gdb/testsuite/gdb.threads/vfork-follow-child-exit.c new file mode 100644 index 00000000000..6ae254cce96 --- /dev/null +++ b/gdb/testsuite/gdb.threads/vfork-follow-child-exit.c @@ -0,0 +1,52 @@ +/* This testcase is part of GDB, the GNU debugger. + + Copyright 2019 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 * +f (void *arg) +{ + int res = vfork (); + + if (res == -1) + { + perror ("vfork"); + return NULL; + } + else if (res == 0) + { + /* Child. */ + _exit (0); + } + else + { + /* Parent. */ + return NULL; + } +} + +int +main (void) +{ + pthread_t tid; + + pthread_create (&tid, NULL, f, NULL); + pthread_join (tid, NULL); + return 0; +} diff --git a/gdb/testsuite/gdb.threads/vfork-follow-child-exit.exp b/gdb/testsuite/gdb.threads/vfork-follow-child-exit.exp new file mode 100644 index 00000000000..f07215d41c6 --- /dev/null +++ b/gdb/testsuite/gdb.threads/vfork-follow-child-exit.exp @@ -0,0 +1,60 @@ +# Copyright (C) 2019 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 following a vfork child that exits, when the vfork parent is a +# threaded program, and it's a non-main thread that vforks. + +standard_testfile + +if {[build_executable "failed to prepare" $testfile $srcfile {debug pthreads}]} { + return -1 +} + +# DETACH indicates whether "set detach-on-fork" is enabled. It is +# either "on" or "off". + +proc test_vfork {detach} { + global binfile + + clean_restart $binfile + + if ![runto_main] then { + fail "can't run to main" + return 0 + } + + gdb_test_no_output "set follow-fork-mode child" + gdb_test_no_output "set detach-on-fork $detach" + + if {$detach == "off"} { + gdb_test "continue" \ + [multi_line \ + "Attaching after .* vfork to child .*" \ + ".*New inferior 2 .*" \ + ".*Inferior 2 .*exited normally.*"] + } else { + gdb_test "continue" \ + [multi_line \ + "Attaching after .* vfork to child .*" \ + ".*New inferior 2 .*" \ + ".*Detaching vfork parent process .* after child exit.*" \ + ".*Inferior 1 .* detached.*" \ + ".*Inferior 2 .*exited normally.*"] + } +} + +foreach_with_prefix detach-on-fork {"off" "on"} { + test_vfork ${detach-on-fork} +} -- 2.14.5