From: Eric Dumazet <dada1@cosmosbay.com>
To: Andre Ben Hamou <andre@bluetheta.com>
Cc: linux-kernel@vger.kernel.org
Subject: Re: Multithread select() bug
Date: Mon, 10 May 2004 23:57:49 +0200 [thread overview]
Message-ID: <409FFADD.7050204@cosmosbay.com> (raw)
In-Reply-To: <409FF38C.7080902@bluetheta.com>
Andre Ben Hamou wrote:
> Here's the scenario...
>
> - parent thread P creates a connected socket pair S[0, 1]
> - P spawns a child thread C and passes it S
> - C selects on S[0]
> - P closes S[0]
>
Your program is racy and have undefined behavior.
A thread should not close a handle 'used by another thread blocked in a
sytemcall'
The race is : if a thread does a close(fd), then the fd value may be
reused by another thread during an open()/socket()/dup()... syscall, and
the first thread could issue the select() syscall (or
read()/write()/...) on the bad file.
Some Unixes defines different semantics (Solaris comes to mind), but
these semantics are not part of standards.
Eric
> As I understand the semantics of the select call, C should now return
> immediately in response to the closure (and it does on Mac OS X).
> However, the following test code behaves otherwise for the two test
> cases I've tried (2.4.21 and 2.6.5). Compilation command used: 'gcc
> foobar.c -lpthread'.
>
> Cheers,
>
> Andre Ben Hamou
> Imperial College London
>
>
> --- BEGIN TEST CODE (foobar.c)---
>
> #include <assert.h> // assert
> #include <pthread.h> // pthread_create
> #include <sys/select.h> // select
> #include <sys/types.h> // socketpair
> #include <sys/socket.h> // socketpair
> #include <unistd.h> // sleep
> #include <stdio.h> // printf
>
> void *threadFuntion (void *sockets) {
> int socket = ((int *)sockets)[0];
> struct timeval timeout = {tv_sec: 5, tv_usec: 0};
>
> // Allocate a file descriptor set with the passed socket
> fd_set fds;
> FD_ZERO (&fds);
> FD_SET (socket, &fds);
>
> // Select to read / register exceptions on the FD set
> select (socket + 1, &fds, NULL, &fds, &timeout);
>
> return NULL;
> }
>
> int main (void) {
> int sockets[2];
> pthread_t thread;
>
> // Create a connected pair of sockets
> assert (socketpair (PF_UNIX, SOCK_STREAM, 0, sockets) != -1);
> printf ("sockets: {%i, %i}\n", sockets[0], sockets[1]);
>
> // Create a POSIX thread
> // - use the default configuration
> // - invoke 'threadFunction' as the root function of the thread
> // - pass the socket array to 'threadFunction'
> assert (pthread_create (&thread,
> NULL,
> threadFuntion,
> sockets) == 0);
>
> // Wait for a second and then close the socket being selected on
> sleep (1);
> assert (close (sockets[0]) == 0);
> printf ("Socket closed\n");
>
> // Wait for the thread to exit - SHOULD BE ~ INSTANTANEOUS
> assert (pthread_join (thread, NULL) == 0);
> printf ("Thread joined\n");
>
> assert (close (sockets[1]) == 0);
> return 0;
> }
>
> --- END TEST CODE ---
>
next prev parent reply other threads:[~2004-05-10 21:59 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2004-05-10 21:26 Multithread select() bug Andre Ben Hamou
2004-05-10 21:47 ` Davide Libenzi
2004-05-10 21:56 ` Andre Ben Hamou
2004-05-10 22:09 ` Davide Libenzi
2004-05-10 21:57 ` Eric Dumazet [this message]
2004-05-10 22:11 ` Andre Ben Hamou
2004-05-10 22:31 ` Eric Dumazet
2004-05-10 23:01 ` Andre Ben Hamou
2004-05-11 6:07 ` Armin Schindler
2004-05-11 6:24 ` Eric Dumazet
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=409FFADD.7050204@cosmosbay.com \
--to=dada1@cosmosbay.com \
--cc=andre@bluetheta.com \
--cc=linux-kernel@vger.kernel.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: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.