]> git.karo-electronics.de Git - karo-tx-linux.git/commit
rbd: kill create_snap sysfs entry
authorAlex Elder <elder@inktank.com>
Fri, 10 Aug 2012 20:12:10 +0000 (13:12 -0700)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Thu, 17 Jan 2013 16:51:20 +0000 (08:51 -0800)
commite4a4a559d2ba0d50059a85967b4c4c707863a65b
treef381d4348d414e9d83cb24a9ee48aca0386abc14
parent3aa540b869b9bd734ec60246ae1dfe35ab3530e0
rbd: kill create_snap sysfs entry

Josh proposed the following change, and I don't think I could
explain it any better than he did:

    From: Josh Durgin <josh.durgin@inktank.com>
    Date: Tue, 24 Jul 2012 14:22:11 -0700
    To: ceph-devel <ceph-devel@vger.kernel.org>
    Message-ID: <500F1203.9050605@inktank.com>
    From: Josh Durgin <josh.durgin@inktank.com>

    Right now the kernel still has one piece of rbd management
    duplicated from the rbd command line tool: snapshot creation.
    There's nothing special about snapshot creation that makes it
    advantageous to do from the kernel, so I'd like to remove the
    create_snap sysfs interface.  That is,
/sys/bus/rbd/devices/<id>/create_snap
    would be removed.

    Does anyone rely on the sysfs interface for creating rbd
    snapshots?  If so, how hard would it be to replace with:

rbd snap create pool/image@snap

    Is there any benefit to the sysfs interface that I'm missing?

    Josh

This patch implements this proposal, removing the code that
implements the "snap_create" sysfs interface for rbd images.
As a result, quite a lot of other supporting code goes away.

[elder@inktank.com: commented out rbd_req_sync_exec() to avoid warning]

Suggested-by: Josh Durgin <josh.durgin@inktank.com>
Signed-off-by: Alex Elder <elder@inktank.com>
Reviewed-by: Josh Durgin <josh.durgin@inktank.com>
(based on commit 02cdb02ceab1f3dd9ac2bc899fc51f0e0e744782)
Documentation/ABI/testing/sysfs-bus-rbd
drivers/block/rbd.c