From owner-freebsd-current Mon Feb 20 15:19:50 1995 Return-Path: current-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.9/8.6.6) id PAA23053 for current-outgoing; Mon, 20 Feb 1995 15:19:50 -0800 Received: from vector.eikon.e-technik.tu-muenchen.de (vector.eikon.e-technik.tu-muenchen.de [192.48.107.36]) by freefall.cdrom.com (8.6.9/8.6.6) with ESMTP id PAA22933 for ; Mon, 20 Feb 1995 15:19:24 -0800 Received: (from jhs@localhost) by vector.eikon.e-technik.tu-muenchen.de (8.6.9/8.6.9) id MAA12218; Sun, 19 Feb 1995 12:58:57 +0100 Date: Sun, 19 Feb 1995 12:58:57 +0100 From: Julian Howard Stacey Message-Id: <199502191158.MAA12218@vector.eikon.e-technik.tu-muenchen.de> To: current@FreeBSD.org Subject: Possible kern.maxproc fatal bug Reply-To: jhs@regent.e-technik.tu-muenchen.de Sender: current-owner@FreeBSD.org Precedence: bulk I report this in case it points a finger towards inadequate limit checking in the kernel: (how i managed to induce the bug, with my mangled ports make set up, doesn't worry me, it's the kernel reaction that is of interest ) As last line in my /etc/rc.local I used to have sysctl -w kern.maxproc=300 An innocent cd /usr/ports ; make -i would then always crash the system (even when compiled as normal user, & with everything in ports having been changed to owner jhs (thus no suid 0 stuff lurking in /usr/ports)). Without the sysctl -w kern.maxproc=300 I would merely get on console: Feb 19 11:38:32 vector kernel.nu: proc: table is full (The system was forking a series of nominal ncftp's & makes recursively) BTW I have: NCFTP=abort /usr/vsl/bin/abort & my abort generates Signal 6 (SIGABRT) I do this as a crude way of forcing make to blunder on & compile most ports, without constantly needing my pay-as-you-use modem slip link. --- Julian Stacey