pthread_cond_timedwait relocks forever


 
Thread Tools Search this Thread
Top Forums Programming pthread_cond_timedwait relocks forever
# 1  
Old 07-02-2010
pthread_cond_timedwait relocks forever

looking in pthread's source code I can see that as an epilogue both pthread_cond_timedwait and pthread_cond_wait will try to relock the mutex by means of __pthread_mutex_cond_lock.

Does this mean that any of them both could eventually block forever if the mutex is never again available after waiting for the conditional variable signal? This does not seem to be the expected behavior for a timed call.

thanks,
Rodrigo
# 2  
Old 07-03-2010
Quote:
looking in pthread's source code I can see that as an epilogue both pthread_cond_timedwait and pthread_cond_wait will try to relock the mutex by means of __pthread_mutex_cond_lock.
Whose pthread source code? There are lots of pthread implementations out there.

Quote:
Does this mean that any of them both could eventually block forever if the mutex is never again available after waiting for the conditional variable signal?
You should acquire the mutex before calling phtread_cond_wait
Code:
pthread_mutex_lock(&m)

/* sleep wait */
while (!condition)
      pthread_cond_wait(&c, &m);

do_your_thing();

pthread_mutex_unlock(&m);

# 3  
Old 07-04-2010
linux, glibc 2.5

Right, before waiting on a conditional variable you must hold the mutex. But once within the 'wait' method itself the mutex will be released while waiting on the conditional variable queue (waiting for the conditional variable being signaled or broadcast.)

Once the signal arrives then 'wait' must guaranty that the mutex is acquired again before returning to the 'wait' caller. That's the basic mechanics for a conditional variable. Reacquiring the mutex, as much as I understand, is always something that one would expect implemented with the API at hand, in this case pthread.

My point is that in this pthread implementation relocking the mutex goes without a timed call, that is, the whole method could block forever if something went wrong with the thread that should have released the mutex. Therefore, it is possible that pthread_cond_timedwait would block for ever, which is exactly what one is trying to avoid by using a timed version of 'wait'.
# 4  
Old 07-04-2010
Quote:
Therefore, it is possible that pthread_cond_timedwait would block for ever, which is exactly what one is trying to avoid by using a timed version of 'wait'.
Yes, it is very possible for that to happen.

I am not sure what your understanding of how pthread_cond_timedwait works is. I would point out that once the wait time expires and is turned off, the thread is moved off the sleep queue. Likewise once a sleeping thread is signaled, even by error, it is taken off the sleep queue and the wait timer is turned off. This does not mean that it re-acquires the mutex, etc. If it takes another week before reacquiring the mutex, too bad. If it never gets the mutex back because the other thread has not released it, too bad.

If you have the Lewis & Berg book on Pthreads, this issue is discussed on pages 94 -99. Dave Butenhof also discusses this issue in his book on Pthreads. See section 3.3.2 et al.

BTW, a thought just struck me, are you trying to port a Win32 application which uses eventing to GNU/Linux?
# 5  
Old 07-04-2010
Quote:
Originally Posted by fpmurphy
Yes, it is very possible for that to happen.

I am not sure what your understanding of how pthread_cond_timedwait works is. I would point out that once the wait time expires and is turned off, the thread is moved off the sleep queue. Likewise once a sleeping thread is signaled, even by error, it is taken off the sleep queue and the wait timer is turned off. This does not mean that it re-acquires the mutex, etc. If it takes another week before reacquiring the mutex, too bad. If it never gets the mutex back because the other thread has not released it, too bad.
i see that your understanding is then the same as mine.

realizing this detail in pthread (at least for the implementation i'm using) was a bit of a surprise. i try to use timed calls (pass timeout) just to be sure that the application will give a last sign of failure before giving up, instead of blocking there silent for ever.

Quote:
Originally Posted by fpmurphy
If you have the Lewis & Berg book on Pthreads, this issue is discussed on pages 94 -99. Dave Butenhof also discusses this issue in his book on Pthreads. See section 3.3.2 et al.

BTW, a thought just struck me, are you trying to port a Win32 application which uses eventing to GNU/Linux?
fortunately no!
Login or Register to Ask a Question

Previous Thread | Next Thread

9 More Discussions You Might Find Interesting

1. UNIX for Dummies Questions & Answers

Forever -w option

I am trying to use the forever command. I can get it to work if I do not use the w option to watch for changes and cause an automatic restart on a change to the contents of the directory being watched. I would really like to use the watch option. Is this option fully implemented? here is an... (4 Replies)
Discussion started by: barrygordon
4 Replies

2. Red Hat

Failed dependencies loop forever

Hello All, I was trying to install one rpm and it failed due to missing dependencies, when I try to look at the dependencies and try to install them it is asking for 100+ dependencies, did any one ever face this problem? how can we fix this? rpm -ivh /var/tmp/erlang-R15B-02.1.el6.x86_64.rpm... (0 Replies)
Discussion started by: lovesaikrishna
0 Replies

3. Shell Programming and Scripting

\n have to make a newline forever?

I'm trying to make a little script, but I have a problem... I'm trying to sed a list of files made with a ls > filename.txt... Two variables (or i may call them constant because they are fixed values :D): ststr1 and ststr2 I want to sed s/"$ststr1"/"$ststr2"/g filename.txt >... (5 Replies)
Discussion started by: maxlamax
5 Replies

4. Programming

How use pthread_cond_timedwait without clock_gettime?

I need to use pthread_cond_timedwait, which is available in my old embedded Linux for PPC. But I don't have clock_gettime... So, I build this replacement for clock_gettime, and it seems to work: #include <sys/time.h> #include <time.h> int clock_gettime_replacement(struct timespec *now) {... (1 Reply)
Discussion started by: fcolombo
1 Replies

5. IP Networking

valid_lft forever preferred_lft forever <-- what does this mean?

Just looking at my ethernet interface.. I see this response... what does this mean...? ipconfig... lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 inet 127.0.0.1/8 scope host lo inet6 ::1/128 scope host valid_lft... (0 Replies)
Discussion started by: jimmyc
0 Replies

6. Linux

I want to mount my disk forever

Hi guys! I've just mounted my drive in fstab: /dev/sdb /myfolder ext3 defaults 0 0 and rebooted linux. I've got severel failers during booting process and also I can't login as root first time: login: root password:root incorrect login login:user password: user ... (1 Reply)
Discussion started by: Junior Admin
1 Replies

7. Solaris

user password forever

Hi I am very new for Solaris, I want to make some users' passwords never expired. My ssytem kernel is: 5.8 # uname -a SunOS sspfs_svr 5.8 Generic_117000-01 sun4u sparc SUNW,Netra-240 Could you make some advice? Thanks (5 Replies)
Discussion started by: xramm
5 Replies

8. Programming

msgrcv pending forever !!!

When I am using msgrcv to get a message from a queue, in case of msgsnd some error, the msgrcv thread will waiting forever. Is there some way that I can specify a time out value for this queue ? just let msgrcv wait for some time, if no message comes during this time slot, msgrcv just return... (3 Replies)
Discussion started by: Yun Gang Chen
3 Replies

9. AIX

AIX 4.3.3 takes forever to log in

Hi all, my RS/6k 7043 150 with aix 4.3.3 takes FOREVER to log in. When I power the machine on, the boot process procedes normally and I get 2 short beeps (which I don't recall hearing before) and then I get the login window. If I log in, as root, say, the machine goes to its usual blue screen... (3 Replies)
Discussion started by: Jwoollard
3 Replies
Login or Register to Ask a Question