From patchwork Tue Dec 16 16:53:54 2014 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Pedro Alves X-Patchwork-Id: 4281 Received: (qmail 22334 invoked by alias); 16 Dec 2014 17:35:25 -0000 Mailing-List: contact gdb-patches-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Unsubscribe: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sourceware.org Delivered-To: mailing list gdb-patches@sourceware.org Received: (qmail 22310 invoked by uid 89); 16 Dec 2014 17:35:22 -0000 Authentication-Results: sourceware.org; auth=none X-Virus-Found: No X-Spam-SWARE-Status: No, score=-2.0 required=5.0 tests=AWL, BAYES_00, SPF_HELO_PASS, SPF_PASS, T_RP_MATCHES_RCVD autolearn=ham version=3.3.2 X-HELO: mx1.redhat.com Received: from mx1.redhat.com (HELO mx1.redhat.com) (209.132.183.28) by sourceware.org (qpsmtpd/0.93/v0.84-503-g423c35a) with (AES256-GCM-SHA384 encrypted) ESMTPS; Tue, 16 Dec 2014 17:35:20 +0000 Received: from int-mx14.intmail.prod.int.phx2.redhat.com (int-mx14.intmail.prod.int.phx2.redhat.com [10.5.11.27]) by mx1.redhat.com (8.14.4/8.14.4) with ESMTP id sBGGs2Sg002323 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-GCM-SHA384 bits=256 verify=FAIL) for ; Tue, 16 Dec 2014 11:54:02 -0500 Received: from brno.lan (ovpn01.gateway.prod.ext.ams2.redhat.com [10.39.146.11]) by int-mx14.intmail.prod.int.phx2.redhat.com (8.14.4/8.14.4) with ESMTP id sBGGrt56018425 for ; Tue, 16 Dec 2014 11:54:01 -0500 From: Pedro Alves To: gdb-patches@sourceware.org Subject: [PATCH 5/5] Test attaching to a program that constantly spawns short-lived threads Date: Tue, 16 Dec 2014 16:53:54 +0000 Message-Id: <1418748834-27545-6-git-send-email-palves@redhat.com> In-Reply-To: <1418748834-27545-1-git-send-email-palves@redhat.com> References: <1418748834-27545-1-git-send-email-palves@redhat.com> Before the previous fixes, on Linux, this would trigger several different problems, like: [New LWP 27106] [New LWP 27047] warning: unable to open /proc file '/proc/-1/status' [New LWP 27813] [New LWP 27869] warning: Can't attach LWP 11962: No child processes Warning: couldn't activate thread debugging using libthread_db: Cannot find new threads: debugger service failed warning: Unable to find libthread_db matching inferior's thread library, thread debugging will not be available. gdb/testsuite/ 2014-12-16 Pedro Alves * gdb.threads/attach-many-short-lived-threads.c: New file. * gdb.threads/attach-many-short-lived-threads.exp: New file. --- .../gdb.threads/attach-many-short-lived-threads.c | 117 ++++++++++++++++++ .../attach-many-short-lived-threads.exp | 135 +++++++++++++++++++++ 2 files changed, 252 insertions(+) create mode 100644 gdb/testsuite/gdb.threads/attach-many-short-lived-threads.c create mode 100644 gdb/testsuite/gdb.threads/attach-many-short-lived-threads.exp diff --git a/gdb/testsuite/gdb.threads/attach-many-short-lived-threads.c b/gdb/testsuite/gdb.threads/attach-many-short-lived-threads.c new file mode 100644 index 0000000..c0bcaf5 --- /dev/null +++ b/gdb/testsuite/gdb.threads/attach-many-short-lived-threads.c @@ -0,0 +1,117 @@ +/* This testcase is part of GDB, the GNU debugger. + + Copyright 2014 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 + +pthread_t main_thread; +pthread_attr_t detached_attr; +pthread_attr_t joinable_attr; + +/* Data passed to joinable threads on creation. This is allocated on + the heap and ownership transferred from parent to child. (We do + this because it's not portable to cast pthread_t to pointer.) */ + +struct thread_arg +{ + pthread_t parent; +}; + +void *thread_fn (void *arg); + +/* Wrapper for pthread_create. */ + +void +create_thread (pthread_attr_t *attr, struct thread_arg *arg) +{ + pthread_t child; + int rc; + + while ((rc = pthread_create (&child, attr, thread_fn, arg)) != 0) + { + fprintf (stderr, "unexpected error from pthread_create: %s (%d)\n", + strerror (rc), rc); + sleep (1); + } +} + +/* Entry point for threads. Detached threads have a NULL parameter. + Joinable threads first join their parent before spawning a new + child (and exiting). The parent's tid is passed as pthread_create + argument. */ + +void * +thread_fn (void *arg) +{ + struct thread_arg *p = arg; + pthread_attr_t *attr; + + if (p != NULL) + { + if (p->parent != main_thread) + assert (pthread_join (p->parent, NULL) == 0); + + attr = &joinable_attr; + p->parent = pthread_self (); + } + else + { + attr = &detached_attr; + } + + create_thread (attr, p); + return NULL; +} + +int +main (int argc, char *argv[]) +{ + int i; + int n_threads = 100; + + if (argc > 1) + n_threads = atoi (argv[1]); + + pthread_attr_init (&detached_attr); + pthread_attr_setdetachstate (&detached_attr, PTHREAD_CREATE_DETACHED); + pthread_attr_init (&joinable_attr); + pthread_attr_setdetachstate (&detached_attr, PTHREAD_CREATE_JOINABLE); + + main_thread = pthread_self (); + + for (i = 0; i < n_threads; ++i) + { + struct thread_arg *p; + + /* Some threads a joinable, others are detached. This exercises + different code paths in the runtime. */ + + p = malloc (sizeof *p); + p->parent = main_thread; + create_thread (&joinable_attr, p); + + create_thread (&detached_attr, NULL); + } + + /* Long enough for all the attach/detach sequences done by the .exp + file. */ + sleep (180); + return 0; +} diff --git a/gdb/testsuite/gdb.threads/attach-many-short-lived-threads.exp b/gdb/testsuite/gdb.threads/attach-many-short-lived-threads.exp new file mode 100644 index 0000000..b14510c --- /dev/null +++ b/gdb/testsuite/gdb.threads/attach-many-short-lived-threads.exp @@ -0,0 +1,135 @@ +# Copyright 2008-2014 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 attaching to a program that is constantly spawning short-lived +# threads. The stresses the edge cases of attaching to threads that +# have just been created or are in process of dying. In addition, the +# test attaches, debugs, detaches, reattaches in a loop a few times, +# to stress the behavior of the debug API around detach (some systems +# end up leaving stale state behind that confuse the following +# attach). + +# Manipulation with PID on target is not supported. +if [is_remote target] then { + return 0 +} + +standard_testfile + +# The test proper. See description above. + +proc test {} { + global binfile + global gdb_prompt + global decimal + + clean_restart ${binfile} + + set testpid [spawn_wait_for_attach $binfile] + + set attempts 10 + for {set attempt 1} { $attempt <= $attempts } { incr attempt } { + with_test_prefix "iter $attempt" { + set attached 0 + set eperm 0 + set test "attach" + gdb_test_multiple "attach $testpid" $test { + -re "new threads in iteration" { + # Seen when "set debug libthread_db" is on. + exp_continue + } + -re "warning: Cannot attach to lwp $decimal: Operation not permitted" { + # On Linux, PTRACE_ATTACH sometimes fails with + # EPERM, even though /proc/PID/status indicates + # the thread is running. + set eperm 1 + exp_continue + } + -re "debugger service failed.*$gdb_prompt $" { + fail $test + } + -re "$gdb_prompt $" { + if {$eperm} { + kfail "gdb/NNNN" "$test (EPERM)" + } else { + pass $test + } + } + -re "Attaching to program.*process $testpid.*$gdb_prompt $" { + pass $test + } + } + + # Sleep a bit and try updating the thread list. We should + # know about all threads already at this point. If we see + # "New Thread" or similar being output, then "attach" is + # failing to actually attach to all threads in the process, + # which would be a bug. + sleep 1 + set saw_new 0 + set test "info threads" + gdb_test_multiple $test $test { + -re "New " { + set saw_new 1 + exp_continue + } + -re "$gdb_prompt $" { + } + } + + gdb_assert !$saw_new "no new threads" + + # Force breakpoints always inserted, so that threads we might + # have failed to attach to hit them even when threads we do + # know about are stopped. + gdb_test_no_output "set breakpoint always-inserted on" + + # Run to a breakpoint a few times. A few threads should spawn + # and die meanwhile. This checks that thread creation/death + # events carry on correctly after attaching. Also, be + # detaching from the program and reattaching, we check that + # the program doesn't die due to gdb leaving a pending + # breakpoint hit on a new thread unprocessed. + gdb_test "break thread_fn" "Breakpoint.*" "break thread_fn" + + # Wait a bit, to give time for most threads to hit the + # breakpoint, including threads we might have failed to + # attach. + sleep 2 + + set bps 3 + for {set bp 1} { $bp <= $bps } { incr bp } { + gdb_test "continue" "Breakpoint.*" "break at thread_fn: $bp" + } + + if {$attempt < $attempts} { + gdb_test "detach" "Detaching from.*" + } else { + gdb_test "kill" "" "kill process" "Kill the program being debugged.*y or n. $" "y" + } + + gdb_test_no_output "set breakpoint always-inserted off" + delete_breakpoints + } + } + + remote_exec target "kill -9 ${testpid}" +} + +if {[prepare_for_testing "failed to prepare" $testfile $srcfile {debug pthreads}] == -1} { + return -1 +} + +test