[Parted-edge] Real partition ID support

David Cantrell dcantrell at redhat.com
Thu Feb 22 23:56:00 UTC 2007


On Fri, 2007-02-23 at 01:43 +0530, Debarshi 'Rishi' Ray wrote:
> http://parted.alioth.debian.org/cgi-bin/trac.cgi/wiki/PlanningEdge
> 
> This is about the "real partition ID support" that we are planning. Based on
> a conversation with Anant and Otavio in #parted, this what I think we need to
> do:
> 
> a. Provide a mechanism for editing a partitions system identifier. From a
>    user's point of view, this is similar to deleting a partition ('rm') and
>    then creating (mkpart) it without the file-system with the required
>    partition ID. The user can be offered a command, say edpart, to do this.

Sounds reasonable.  Is the editing mechanism only for the partition ID
or will it also be a synonym for resize or move since "editing" could
also include partition boundaries since those are technically properties
of the partition.

My thought is that we shouldn't do that.  We should keep with the
command set we have now in parted(8) and make sure the new command
(either edpart or something else) is restricted to editing values except
the partition boundaries and sizes.

Forgive me if this was already discussed in IRC.  I've been absent from
IRC and today I wasn't working anyway.  Tomorrow I'm offline until the
evening (UTC+1).

> b. Modify the mkpart command to accept partition IDs too.

Good idea.

> c. We already have means to read the partition ID while probing a partition.
>    eg., since extended partitions on MS-DOS disk labels do not have any file
>    system they are identified on the basis of the partition ID only. Study
>    about the partition ID standards and the way the 't' command works in
>    FDisk (util-linux).

Another thing to think about is GPT and the requirements for IDs there.
The hal code:

	http://www.freedesktop.org/wiki/Software_2fhal

Is an interesting project that makes use of libparted, but fills in the
gaps where we don't provide the functionality we need.  Their code that
uses libparted is decently documented as to what we lack in libparted,
so it's worth a look.

But they also have nice code for generating the GPT IDs according to
spec.  The main developer of hal is David Zeuthen, who also works at Red
Hat and sits a few cubicles away from me.

-- 
David Cantrell <dcantrell at redhat.com>
Red Hat / Westford, MA
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : http://lists.alioth.debian.org/pipermail/parted-edge/attachments/20070222/ab68b00d/attachment.pgp


More information about the Parted-edge mailing list