Journals and jscan

[Steve O'Hara-Smith]

	Primarily to get this in the archives for later scrutiny.

On Mon, 22 Jan 2007 11:26:42 -0800 (PST)
Matthew Dillon <dillon at apollo.backplane.com> wrote:

> :	That (and variations on the theme) all produce the same Bad path:
> :messages and paths with leading /s appear in the jscan -d output too. The
> :paths aren't really absolute they are relative to the mount point but
> have :a leading / and so look absolute.
> :
> :	I'm using a very up to date Preview build BTW.
> 
>     Since journaling is still highly experimental I am not going to worry
>     about it for this release, but please remind me after the release and
>     I will look into the bad path problem.

	The leading / is originating in jrecord_write_path in vfs_journal.c,
I added some debug which revealed that the chain of namecache structures
ends with an extra one with nc_nlen set to 0. I have no idea if this extra
record is vital for some purpose so I've added the attached patch to make
jrecord_write_path ignore it rather than try and track down where it is
coming from and remove it (I'd hate to destabilise the namecache stuff -
even in a vkernel).

	Unfortunately it's not enough to make the mirror work - something
seems to be astray in jscan - I'll look into that next.

	PS: The vkernel is *wonderful* :)

-- 
C:>WIN                                      |   Directable Mirror Arrays
The computer obeys and wins.                | A better way to focus the sun
You lose and Bill collects.                 |    licences available see
                                            |    http://www.sohara.org/
Attachment:
vfs_journal.c.patch
-------------- next part --------------
A non-text attachment was scrubbed...
Name: bin00000.bin
Type: application/octet-stream
Size: 753 bytes
Desc: "Description: Binary data"
URL: <http://lists.dragonflybsd.org/pipermail/users/attachments/20070124/4994d012/attachment-0021.bin>