Date: Sun, 1 Oct 2000 13:11:26 +0930 From: Greg Lehey <grog@lemis.com> To: Terry Lambert <tlambert@primenet.com> Cc: "Karsten W. Rohrbach" <karsten@rohrbach.de>, Andre Albsmeier <andre@akademie3000.de>, Marc Tardif <intmktg@CAM.ORG>, freebsd-fs@FreeBSD.ORG, freebsd-scsi@FreeBSD.ORG Subject: Partitioning (was: ccd with other filesystems) Message-ID: <20001001131126.L43885@wantadilla.lemis.com> In-Reply-To: <200010010248.TAA19148@usr05.primenet.com>; from tlambert@primenet.com on Sun, Oct 01, 2000 at 02:48:53AM %2B0000 References: <20001001120453.I43885@wantadilla.lemis.com> <200010010248.TAA19148@usr05.primenet.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sunday, 1 October 2000 at 2:48:53 +0000, Terry Lambert wrote:
>>> The kernel and other files loaded by the loader myst be below
>>> the 1024 cylinder boundary on the disk, since the FrrBSD boot
>>> loader, unlike Linux's LILO, can not read past cylinder 1024,
>>> since it does not understand how to make LBA BIOS calls
>>> properly.
>>
>> This is no longer the case. The restriction was lifted a few months
>> back.
>
> Someone should tell my Sony Vaio, which hated having 4.1 installed
> on it past 8G on an 18G drive.
>
>>> I have several systems, where the entirety of the disk (the "c"
>>> partition) is mounted as a single file system.
>>
>> You can do it, but it's not a good idea. I'd like to see a good
>> reason for doing this.
>
> If "c" is defined to be "the whole disk" , and you want to use "the
> whole disk", it makes sense.
No, you don't *ever* want a UFS on the whole disk, you want it on a
partition, because those are the objects on which file systems work.
The partition may well cover the entire disk, but that doesn't give it
the same semantics as an unused partition which covers it.
> I uses to mount most of my CDROMs that way, as well.
That's OK. CD-ROMs don't have partitions. It's probably a bug that
we even have partition letters for non-partitioned media.
>>> This appears to be a problem with not checking the label for
>>> overlap, since a mounted FS should not be spam'able under any
>>> circumstances. Protecting people from spam'ming unmounted FSs by
>>> pounding on "c" might be a laudable goal, but provides only a tiny
>>> amount of additional protection.
>>
>> This is a separate issue. Yes, disklabel should warn about a number
>> of things, including overlapping partitions and incorrect partition
>> types (c should be "unused", because by definition it overlaps all
>> other partitions).
>
> I really _don't_ want the use of "c" broken ("fixed").
>
> The "c" partition is not "defined" to overlap; it merely does so by
> default and by istorical convention.
Agreed. But there were good reasons for this.
> I threw away this convention on many of my systems long ago, when I
> resigned myself to aving a DOS parititon table on my machines, when
> the Alpha and PReP platforms decided to require it as well.
Is your 'h' key sticking? :-)
I strongly object to the Microsoft "partition" table, and I don't use
it myself. And of course you're welcome to use whatever you find
convenient. It's not until you advocate making this a standard way
that anybody can have any objection.
Greg
--
Finger grog@lemis.com for PGP public key
See complete headers for address and phone numbers
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-scsi" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20001001131126.L43885>