From owner-freebsd-commit  Wed Oct  4 18:15:40 1995
Return-Path: owner-commit
Received: (from root@localhost)
          by freefall.freebsd.org (8.6.12/8.6.6) id SAA12390
          for freebsd-commit-outgoing; Wed, 4 Oct 1995 18:15:40 -0700
Received: (from root@localhost)
          by freefall.freebsd.org (8.6.12/8.6.6) id SAA12378
          for cvs-all-outgoing; Wed, 4 Oct 1995 18:15:37 -0700
Received: (from root@localhost)
          by freefall.freebsd.org (8.6.12/8.6.6) id SAA12367
          for cvs-sys-outgoing; Wed, 4 Oct 1995 18:15:34 -0700
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.34])
          by freefall.freebsd.org (8.6.12/8.6.6) with ESMTP id SAA12338
          ; Wed, 4 Oct 1995 18:15:20 -0700
Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.9/8.6.9) id LAA23044; Thu, 5 Oct 1995 11:10:27 +1000
Date: Thu, 5 Oct 1995 11:10:27 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199510050110.LAA23044@godzilla.zeta.org.au>
To: CVS-commiters@freefall.freebsd.org, bde@freefall.freebsd.org,
        cvs-sys@freefall.freebsd.org
Subject: Re: cvs commit: src/sys/net if_ppp.c
Sender: owner-commit@FreeBSD.org
Precedence: bulk

>  Modified:    sys/net   if_ppp.c
>  Log:
>  Don't wait for output to drain in pppclose().  Discard output immediately
>  for the same reasons as in slclose().
>
>  [stuff that shouldn't go in 2.1]

This should go in 2.1.

Unlike for slip, this only helps when pppd is killed without giving it a
chance to clean up.  pppd uses a waiting form of tcsetattr() to restore
the tty state as part of cleaning up, and it doesn't bother to switch
the line discipline back to TTYDISC before calling tcsetattr(), so it
hangs before pppclose() is called.  This is not as serious as hanging
in close() in exit(), because not all signals are blocked.  Hanging
pppd's can be killed by sending them 2 SIGTERMs after the signal that
made it start cleaning up (SIGHUPs are blocked).

Bruce