Re: [RFA/testsuite] backtrace past pthread_cond_wait()

[Daniel Jacobowitz <drow@false.org>]

On Wed, Nov 19, 2003 at 11:34:58PM -0800, Joel Brobecker wrote:
> Hello,
> 
> Re: http://sources.redhat.com/ml/gdb-patches/2003-10/msg00440.html
> 
> Here is a new test that would verify that GDB is able to unwind past
> pthread_cond_wait(). On my laptop, I get all PASS, while on a RH9.0
> machine I get one FAIL.

On Debian unstable I get all PASS, with both LinuxThreads and NPTL. 
Not bad.  Didn't even have to use the system's separate debug info
(I keep the debug info in /usr/lib/debug and my test GDB tree pointed
at /usr/local/lib/debug, usually).

> I am not too familiar with pthreads, so I hope my program is
> portable. In particular, I am worried about cygwin... Some guidance
> would be much appreciated (in Ada, we use tasks, which syntax and
> behavior is defined by the Ada Reference Manual, so the issue about
> portability is left to the compiler/runtime developers).
> 
> 2003-11-19  J. Brobecker  <brobecker@gnat.com>
> 
>         * gdb.threads/pt.c: New file.
>         * gdb.threads/pthread_cond_wait.exp: New test.
> 
> OK to apply?

My only complaint is about pt.c - please use a more descriptive name.
The test itself looks OK.  I would be happier if you didn't play games
with info threads; the order in which GDB will discover threads is
actually somewhat random using gdbserver, since new threads are not
notified.  But this should be OK anyway.

Michael Snyder is the listed maintainer of gdb.threads.  Michael?

> 
> Thanks,
> -- 
> Joel

> /* A small multi-threaded test case.
> 
>    Copyright 2003
>    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 2 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, write to the Free Software
>    Foundation, Inc., 59 Temple Place - Suite 330,
>    Boston, MA 02111-1307, USA.  */
> 
> #include <pthread.h>
> #include <stdio.h>
> #include <time.h>
> 
> void
> cond_wait (pthread_cond_t *cond, pthread_mutex_t *mut)
> {
>   pthread_mutex_lock(mut);
>   pthread_cond_wait (cond, mut);
>   pthread_mutex_unlock (mut);
> }
> 
> void
> noreturn (void)
> {
>   pthread_mutex_t mut;
>   pthread_cond_t cond;
> 
>   pthread_mutex_init (&mut, NULL);
>   pthread_cond_init (&cond, NULL);
> 
>   /* Wait for a condition that will never be signaled, so we effectively
>      block the thread here.  */
>   cond_wait (&cond, &mut);
> }
> 
> void *
> forever_pthread (void *unused)
> {
>   noreturn ();
> }
> 
> void
> break_me (void)
> {
>   /* Just an anchor to help putting a breakpoint.  */
> }
> 
> int
> main (void)
> {
>   pthread_t forever;
>   const struct timespec ts = { 0, 10000000 }; /* 0.01 sec */
> 
>   pthread_create (&forever, NULL, forever_pthread, NULL);
>   for (;;)
>     {
>       nanosleep (&ts, NULL);
>       break_me();
>     }
> 
>   return 0;
> }
> 

> # Copyright (C) 2003 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 2 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, write to the Free Software
> # Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  
> 
> # Please email any bugs, comments, and/or additions to this file to:
> # bug-gdb@gnu.org
> 
> # This file was written by Joel Brobecker.
> #
> # It verifies that GDB is able to compute a backtrace for a thread
> # being blocked on a call to pthread_cond_wait(). 
> 
> if $tracelevel then {
> 	strace $tracelevel
> }
> 
> set testfile "pt"
> set srcfile ${testfile}.c
> set binfile ${objdir}/${subdir}/${testfile}
> 
> if {[gdb_compile_pthreads "${srcdir}/${subdir}/${srcfile}" "${binfile}" executable [list debug "incdir=${objdir}"]] != "" } {
>     return -1
> }
> 
> gdb_exit
> gdb_start
> gdb_reinitialize_dir $srcdir/$subdir
> gdb_load ${binfile}
> 
> gdb_test "break break_me" \
>          "Breakpoint 1 at .*: file .*pt.c, line .*" \
>          "breakpoint on break_me"
> 
> gdb_test "run" \
>          ".*Breakpoint 1, break_me ().*" \
>          "run to break_me"
> 
> # Depending on the thread implementation (Linux Threads, NPTL, etc...),
> # The number of threads known to GDB may vary. We should have at least
> # two threads, but we may have more. We want to switch to the pthread
> # we created via pthread_create, and its ID should be the highest ID
> # of all threads known to GDB.
> set thread_id ""
> send_gdb "info threads\n"
> gdb_expect {
>     -re "\s*(\[0-9\]+) Thread.*$gdb_prompt $" {
>         set thread_id $expect_out(1,string)
>     }
>     default {
>         fail "Unable to find thread id, defaulting to 2"
>         set thread_id "2"
>     }
> }
>     
> 
> # Switch to thread 2, which is doing a forever wait...
> gdb_test "thread $thread_id" \
>          ".*Switching to thread $thread_id.*" \
>          "switching to thread $thread_id"
> 
> # Verify that we are able to get a sensible backtrace, including
> # the frame for the pthread_cond_wait() call.
> gdb_test "backtrace" \
>          ".*pthread_cond_wait.*cond_wait.*noreturn.*forever_pthread.*" \
>          "backtrace in blocked thread"
> 


-- 
Daniel Jacobowitz
MontaVista Software                         Debian GNU/Linux Developer