From owner-freebsd-current Tue Sep 8 15:50:28 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id PAA00279 for freebsd-current-outgoing; Tue, 8 Sep 1998 15:50:28 -0700 (PDT) (envelope-from owner-freebsd-current@FreeBSD.ORG) Received: from highwind.com (hurricane.highwind.com [209.61.45.50]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id PAA00273 for ; Tue, 8 Sep 1998 15:50:24 -0700 (PDT) (envelope-from info@highwind.com) Received: (from info@localhost) by highwind.com (8.8.6/8.8.6) id SAA10366; Tue, 8 Sep 1998 18:50:17 -0400 (EDT) Date: Tue, 8 Sep 1998 18:50:17 -0400 (EDT) Message-Id: <199809082250.SAA10366@highwind.com> From: HighWind Software Information To: rotel@indigo.ie CC: freebsd-current@FreeBSD.ORG In-reply-to: <199809082215.XAA01106@indigo.ie> (message from Niall Smart on Tue, 8 Sep 1998 23:15:53 +0000) Subject: Re: Thread Problems Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG > Frankly, this looks like the fd is not async. The only potentially > blocking call should be select(). But it's the thread library's responsability not to call accept until it is sure that the call will not block.... Remember, on the first call to accept(), libc_r WILL make the real accept() call. It'll do it, however, on a non-blocking fd, realize it got a EWOULDBLOCK, and then put the fd on the list for a later accept() once select() fires off. Could fork()/exec() be involved? We do call fork() and exec(). I strongly suspect the fork() or exec() logic might be buggy and since it sets things back to "blocking", it might effect existing fd's. -Rob To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message