Edgar Toernig <firstname.lastname@example.org> writes:
> Alan Cox wrote:
> > > We have code that uses an alarm() to routinely break ouf of a socket accept
> > > and then perform some processing (check a flag) and then loop back around
> > > to wait for a connection again. The code works fine on Solaris, HP-UX,,
> > > AIX, and other flavours but not on RedHat 6.0.
> > Ugly way to do it (non blocking connect and select() is cleaner)
> > > I noticed that by default, SIGALRM is ignored so I unblocked it with
> > > sigprocmask but we are still not getting the behaviour we require.
> > > Is there something I am missing here?
> > Are you using sigaction to force the signal to interrupt connect and not
> > be restartable ?
> In a similar case I noticed that SA_RESTART is ignored for select().
> IMO a very strange behaviour! Is the xview clock binary really the
> only reason???
If select was restarted it would wait its timeout again.
With a SIGALRM timer in just the right period you could make a select()
block forever, even if it was supposed to timeout. This happens in practice,
e.g. with some versions of the DNS resolver when you have a periodic timer
running. Fixing this would complicate the system call restart mechanism a lot
and also probably break some programs.
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to email@example.com
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Thu Mar 23 2000 - 21:00:21 EST