[sslh] sslh and openbsd

[Yves Rutschle]

On Tue, May 18, 2010 at 12:25:43PM +0200, François FRITZ wrote:
> He submitted a patch to you yesterday and you didn't like it :-).

I object to this caracterisation :p It's just that the
behaviour we're seeing is contradictory with OpenBSD
documentation and I like to go to the bottom of things
before fixing them (to be sure I am fixing the right thing).

> The SA_NOCLDWAIT flag is not propagated to the child process.

Ok, that's actually exactly what I suspected.

Do we agree that it's still in contradiction with OpenBSD
documentation? (from
http://www.rocketaware.com/man/man2/sigaction.2.htm:
  "After a fork(2) or vfork(2),  all signals, the signal
mask, the signal stack, and the restart/interrupt flags are
inherited by the child.")
If it is, then that's a bug in OpenBSD (or in the
documentation) and we should ideally report it with your
test program.


> Here is the patch in the diff format:
> 462a463,464
> >    if (fork() > 0) exit(0); /* Detach */
> >
> 467,468d468
> <    if (fork() > 0) exit(0); /* Detach */

I do like that a lot better ;)

I'll test this and include it in the next release.

Cheers,
Y.