Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 16 Dec 1998 11:55:36 +1030
From:      Greg Lehey <grog@lemis.com>
To:        mjacob@feral.com
Cc:        freebsd-current@FreeBSD.ORG
Subject:   Re: Tape Driver Changes Proposed: Tape Early Warning Behaviour
Message-ID:  <19981216115536.W15815@freebie.lemis.com>
In-Reply-To: <Pine.LNX.4.04.9812141927130.2563-100000@feral-gw>; from Matthew Jacob on Mon, Dec 14, 1998 at 07:31:15PM -0800
References:  <19981215135448.B15815@freebie.lemis.com> <Pine.LNX.4.04.9812141927130.2563-100000@feral-gw>

next in thread | previous in thread | raw e-mail | index | archive | help
On Monday, 14 December 1998 at 19:31:15 -0800, Matthew Jacob wrote:
>>
>> Well, you don't have to go overboard.  Combine compression and
>> density.  We have BSD semantics, I suppose, so the only other thing is
>> no rewind, which we already cater for.  So for, say, an Exabyte
>> 8505XL, you'd have:
>>
>>   /dev/rst0l		8202 mode, no compression
>>   /dev/rst0n		8202 mode, compression
>>   /dev/rst0h		8205 mode, no compression
>>   /dev/rst0c		8205 mode, compression
>>   /dev/nrst0l		8202 mode, no compression, no rewind
>>   /dev/nrst0n		8202 mode, compression, no rewind
>>   /dev/nrst0h		8205 mode, no compression, no rewind
>>   /dev/nrst0c		8205 mode, compression, no rewind
>>
>> Sure, it's more than now, but it shouldn't confuse people too much.
>
> Cool, but there are four possible densities in the current data
> structures. And speeds (has anyone actually ever found a drive that
> *uses* these?). And multiple compression algorithms to select from.
>
> I'm kind of inclined to think that the compromise of
>
> 	Rew/Norew X Compress/Nocompress X Low/High density
>
> is sufficient as long as you can establish that the latter two
> categories are a persistent (not through reboot, tho) cache of possible
> values that you can set via the mt(1) command.

I can't see any particular reason to restrict the minor number format
if it's not necessary.  So we have 4 densities, n (<=4?) speeds and
compression.  That makes 5 bits.  Then we have non-rewind, a maximum
of 16 units per drive and (on a PC) probably not more than 16
controllers.  A total of 14 bits of minor number out of the 24
available: in other words, there should be no problem finding a minor
number format which fits.

The naming is a different matter.  The `convenient' format doesn't
have to cover every possibility; we could think up a `complete' format
at a later date, something like

  /dev/tape/nrc0u5d3s1c

meaning no rewind, raw, controller 0, unit 5, density setting 3, speed
setting 1, compression.  There's no reason to think these out at this
stage, since they're probably not needed; but if we create a minor
number format which can't represent one or the other of them, we're
bound to find a case where it's needed.

Greg
--
See complete headers for address, home page and phone numbers
finger grog@lemis.com for PGP public key

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19981216115536.W15815>