Where to start?

[Peter Jeremy]

On Mon, 2007-Jan-22 10:29:18 -0800, perryh at pluto.rain.com wrote:
>I would guess that if a way could be found to preallocate the
>journal space (as with mkfile(8) in sufficiently-old systems),

mkfile(8)ing a journal is easy.  This would not guarantee that the
journal was a contiguous block though and the journalling code would
also need to be able to follow the journal contents through a block
list chain defined by an inode - this is not difficult but not as easy
as having a single contiguous chunk of space.

>and then record its location in a reasonably-secure location
>(the superblock?), it could be accessed during recovery without
>reference to possibly-corrupt filesystem metadata.

The superblock is the logical location.  There are a number of
spare fields in the superblock that could potentially be used to
contain a journal location.

Files within UFS are described by an inode number so the 'location'
of the journal would be an inode number.  The journal code would
need to verify that the given inode was internally consistent before
it accessed the data.

-- 
Peter Jeremy
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 187 bytes
Desc: not available
Url : http://lists.freebsd.org/pipermail/freebsd-hackers/attachments/20070123/443cad96/attachment.pgp