From owner-freebsd-current Fri Mar 8 15:17:50 1996 Return-Path: owner-current Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id PAA21348 for current-outgoing; Fri, 8 Mar 1996 15:17:50 -0800 (PST) Received: from who.cdrom.com (who.cdrom.com [192.216.222.3]) by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id PAA21293 for ; Fri, 8 Mar 1996 15:17:36 -0800 (PST) Received: from Sisyphos (Sysiphos.MI.Uni-Koeln.DE [134.95.212.10]) by who.cdrom.com (8.6.12/8.6.11) with SMTP id FAA24908 for ; Fri, 8 Mar 1996 05:15:39 -0800 Received: by Sisyphos id AA09086 (5.67b/IDA-1.5 for freebsd-current@FreeBSD.ORG); Fri, 8 Mar 1996 14:15:18 +0100 Message-Id: <199603081315.AA09086@Sisyphos> From: se@zpr.uni-koeln.de (Stefan Esser) Date: Fri, 8 Mar 1996 14:15:17 +0100 In-Reply-To: J Wunsch "Re: New kernel option proposed.." (Mar 8, 9:15) X-Mailer: Mail User's Shell (7.2.6 alpha(2) 7/9/95) To: joerg_wunsch@uriah.heep.sax.de (Joerg Wunsch) Subject: Re: New kernel option proposed.. Cc: freebsd-current@FreeBSD.ORG (FreeBSD-current users) Sender: owner-current@FreeBSD.ORG X-Loop: owner-current@FreeBSD.ORG Precedence: bulk On Mar 8, 9:15, J Wunsch wrote: } Subject: Re: New kernel option proposed.. } As Jordan K. Hubbard wrote: } } > Stefan Esser would like an option which he can use to conditionalize } > code to be "conservative" in the case where you're first installing. } } Too many people run the GENERIC kernel as their normal one. I don't } have an idea which amount of code will be affected, but i'd rather } vote for a sysctl variable (perhaps defaulting to ``conservative'' in } the GENERIC kernel, and being turned off in /etc/rc later). I only want the FAILSAFE define for boot init code. The usage should be documented, and this is what I want it for currently: 1) NCR: No **active** synch. transfer negotiation for CDROM drives, which currently fails with the Chinon 525 and 535. 2) NCR: Have a default of NO tags for disk drives. This can be raised in the running system using "ncrcontrol". 3) NCR: use a default of 1 for MAX_LUN instead of the value of 8 in current kernels. 4) PCI: Possibly use a lower burst length limit. Documenting these effects of FAILSAVE being defined is VERY important, or nobody will understand why a custom kernel fails, when the GENERIC kernel worked. Regards, STefan -- Stefan Esser, Zentrum fuer Paralleles Rechnen Tel: +49 221 4706021 Universitaet zu Koeln, Weyertal 80, 50931 Koeln FAX: +49 221 4705160 ============================================================================== http://www.zpr.uni-koeln.de/~se