From owner-freebsd-current Mon Nov 15 9:15:51 1999 Delivered-To: freebsd-current@freebsd.org Received: from gndrsh.dnsmgr.net (GndRsh.dnsmgr.net [198.145.92.4]) by hub.freebsd.org (Postfix) with ESMTP id 5735514E9A for ; Mon, 15 Nov 1999 09:15:42 -0800 (PST) (envelope-from freebsd@gndrsh.dnsmgr.net) Received: (from freebsd@localhost) by gndrsh.dnsmgr.net (8.9.3/8.9.3) id JAA15320; Mon, 15 Nov 1999 09:15:22 -0800 (PST) (envelope-from freebsd) From: "Rodney W. Grimes" Message-Id: <199911151715.JAA15320@gndrsh.dnsmgr.net> Subject: Re: FreeBSD 4.0 SCSI Tape Driver In-Reply-To: from Matthew Jacob at "Nov 15, 1999 03:37:54 am" To: mjacob@feral.com Date: Mon, 15 Nov 1999 09:15:21 -0800 (PST) Cc: current@FreeBSD.ORG X-Mailer: ELM [version 2.4ME+ PL54 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG > The design phase for FreeBSD 4.0 is coming to a close. There are a couple > of things I'm planning on (belatedly) for the SCSI tape driver. I'd like > feedback and suggestions about these and other things, so pass 'em my way. > > One change I'm thinking about is probably controversial, so I'd like to > get some feedback on it now. Since this is a major release step, this > would be the time to make such changes if at all. > > I'd like to make the *default* tape EOT handling behaviour such that all > tapes use only *one* filemark at EOT rather than the current *two* > filemarks at EOT (except for QIC). Probably one of the highest breakage > items for this driver is someone adding yet another unknown QIC-like tape > device which behaves unhappily when the driver tries to write two > filemarks at the end of tape. Insofar as I know, the convention for two > filemarks at the end of tape is useful only for devices that cannot > determine physical eot (1/2" Reel tapes)- and I haven't seen those around > for quite some time. There seems to be a great amount of confusion about the 2 EOF marks on tapes. It has nothing to do with physical EOT, even the 556BPI 1/2" tape drives on an IBM 1401 can detect physical EOT. The problem is with LOGICAL EOT, most tape drives do not have a logical EOT write command, even modern drives. So when you overwrite a tape how do you tell that you have gotten to the logical end of data, well, you write 2 EOF marks. The other thing that causes lots of folks confusion here is that some tape drives backspace over an EOF mark that is written, thus it gets real fun to put 2 EOF marks on the tape. You have to mt eof, mt fsf, mt eof. Since you do not point out how we are suppose to detect logical EOT on a tape I object to any elimination of dual EOF to indicate logical EOT. > > There already is an ioctl (and control via mt(1)) to change the default > eot model. There could very well also be a config option too. I'd like to > make the 1 Filemark at EOT the default though. I'll have to fix tcopy, > and I want to give some thought so that there are no compatibility > and interchange problems, but if those concerns are adequately covered I > think this is the right thing to do. 1 filemark can not be used for EOT, it is EOF, you can't tell if what you read next is another file or not that may have been left by a previosly longer usage on the tape. > > So- let me know, either via this list or privately. > Thanks in advance... Won't work, or would you care to explain how we are now suppose to detect logical EOT? -- Rod Grimes - KD7CAX @ CN85sl - (RWG25) rgrimes@gndrsh.dnsmgr.net To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message