Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 24 Nov 2006 21:03:02 +0100
From:      Ulf Lilleengen <lulf@stud.ntnu.no>
To:        freebsd-geom@freebsd.org
Cc:        freebsd-current@freebsd.org
Subject:   Gvinum patches for testing
Message-ID:  <20061124200302.GA3482@vimes.studby.ntnu.no>

next in thread | raw e-mail | index | archive | help
Hi,

I'm releasing some patches to gvinum for testing. I'd appriciate any feedback
people can give me by testing them. I'm trying to split my patches as much as I
can, as well as providing a complete patch for most stuff.

All (at least most), of my work can be found here:
http://folk.ntnu.no/lulf/patches/freebsd/gvinum

IMO, the most ready of these are the implementation of setstate for
volumes and plex, the stop command, and attach/detach commands (the latter is
dependant on the first, and includes it, so if you just want to test only
setstate, you can do that), so I'd appreciate testing on these the most right
now, but I welcome everything :)

To apply them just do:

# cp gvinum_setstate.diff /usr/src
# patch < gvinum_setstate.diff

And you should be good to go.

Just replace the gvinum_setstate.diff with any other patch you want to apply.
gvinum_attach_detach.diff for the attach/detach + the setstate code.

I have an implementation of the dumpconfig option as well, but I do not really
like the way I've implemented it, so there will most certainly be changes to
that.

Other than that, I have put my drivegroup-implementation a bit on hold,
until I've completed all old vinum commands first since I feel that to be most
important right now. This means my fully functional implementation of the
'concat' command (although I discovered vinum actually behaves slightly
differently in this matter), will have to be changed a bit. I noticed my
implementation uses pre-created gvinum drives to create volumes, instead of
taking the actual device as a parameter, so I'm currently working on adapting
the code for that, although it's not easy to do it because gvinum is radically
different in some ways compared to vinum. Anyway, all my work on drivegroup can
be found in the gvinum_all_current.diff. Always nice with testing there too,
but a lot can change there.

The next step after the concat command will be the mirror and stripe commands,
which should be fairly easy when I've implemented the functions to be used by
the concat command. I will send a notice out to mailing lists as soon as I've
done more work in these areas (I'm in an exam period at the university right
now, so some time will have to go away to school :) ).

I also have some bugs in gvinum in general that I've not found time to look at,
but I hope to be able to soon.

And at last, it seems grog@ is kind of busy these days, so I have again no
committer to review my changes, but I hope to get some more help from others to
test it at least. Still, it would be nice if I could grab other available
committers attention. Again, I welcome suggestions and complaints.

PS. I also have a style patch, but that shouldn't need any testing :)

-- 
Ulf Lilleengen



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20061124200302.GA3482>