Re: sendto generates SIGSEGV

From: Matti Aarnio (matti.aarnio@zmailer.org)
Date: Sun Jul 29 2001 - 17:30:11 EST


On Sun, Jul 29, 2001 at 11:27:29PM +0200, Oliver Kowalke wrote:
> From: oliver.kowalke@t-online.de (Oliver Kowalke)
> To: linux-kernel@vger.kernel.org
> Subject: sendto generates SIGSEGV
> Date: Sun, 29 Jul 2001 23:27:29 +0200

  Two hours ago you asked this very same question at the linux-net -list.
  At how many lists you are asking this ?

> Hi,
>
> I've written an c++ class which wrappes a socket.. With function writen()
> you can write to an TCP and UDP socket ( TCP : pointer to stuct sockaddr ==
> NULL; UDP : pointer to struct sockaddr != NULL). If I use this function
> with TCP sockets all works. If I call this function for an UDP socket it
> generates an SIGSEGV signal.
> What is wrong?

    You are presenting incomplete source, and worse, incomplete
  debug trace. Furthermore, this really does not look like KERNEL
  problem, which makes it Out-of-Topic issue at these lists.

  From certain signature features in the very scant debug data, I would
  venture to think that you are running it at Solaris, and debugging
  it there.

  Do yourself a favour and learn to use debuggers, and compile for
  debugging. ``Backtrace'' or ``where'' (or whatever the debugger
  of your choice calls it) is your friend in most cases.

  Syscall trace may give you also some clues in some cases.
  At Solaris see: 'man truss'. At Linux see: 'man strace'.

  If you can show strace output with SIGSEGV during a network syscall,
  now THAT would be most appropriate at the linux-net list.

> with regards,
> Oliver

    /Matti Aarnio

> ssize_t writen( const void * vptr, size_t n, struct sockaddr * to)
> {
> size_t nleft = n;
> ssize_t nwritten;
> const char *ptr = static_cast< const char * >( vptr);
>
> struct sigaction new_sa;
> struct sigaction old_sa;
>
> new_sa.sa_handler = SIG_IGN;
> ::sigemptyset( & new_sa.sa_mask);
> new_sa.sa_flags = 0;
> ::sigaction( SIGPIPE, & new_sa, & old_sa);
>
> while ( nleft > 0)
> {
>
> if ( ( nwritten = ::sendto( m_handle, ptr, nleft, 0, to, sizeof * to) )<=0)
> {
> if ( errno == EINTR)
> nwritten = 0; // and call sendto() again
> else if ( errno == EPIPE)
> {
> ::sigaction( SIGPIPE, & old_sa, 0);
>
> return -1; // write to sock with no readers (peer has the sock closed)
>
> }
> else
> {
> ::sigaction( SIGPIPE, & old_sa, 0);
> throw; // error
> }
> }
>
> nleft -= nwritten;
> ptr += nwritten;
> }
> ::sigaction( SIGPIPE, & old_sa, 0); // set to its previous action
>
> return n;
> }
>
> here the output of the generated core file:
>
> Core was generated by 'lt-ogx 192.168.1.3 10000'.
> Program terminated with signal 11, Segmentation fault.
> rw_common (): write: Success.
> warning: unable to get global thread event mask
> [New Thread 1024 (LWP 32350)]
> rw_common (): write: Success.
> warning: stop_or_attach_thread: generic error
> #0 0x0 in ?? ()
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Tue Jul 31 2001 - 21:00:41 EST