Re: [RFC PATCH 18/35] libceph, rbd: Convert some page arrays to ceph_databuf

[David Howells <dhowells@redhat.com>]

Viacheslav Dubeyko <Slava.Dubeyko@ibm.com> wrote:

> >  	/*
> >  	 * The response data for a STAT call consists of:
> > @@ -2118,14 +2118,12 @@ static int rbd_osd_setup_stat(struct ceph_osd_request *osd_req, int which)
> >  	 *         le32 tv_nsec;
> >  	 *     } mtime;
> >  	 */
> > -	pages = ceph_alloc_page_vector(1, GFP_NOIO);
> > -	if (IS_ERR(pages))
> > -		return PTR_ERR(pages);
> > +	dbuf = ceph_databuf_reply_alloc(1, 8 + sizeof(struct ceph_timespec), GFP_NOIO);
> 
> What this 8 + sizeof(struct ceph_timespec) means? Why do we use 8 here? :)

See the comment that's partially obscured by the patch hunk line:

	/*
	 * The response data for a STAT call consists of:
	 *     le64 length;
	 *     struct {
	 *         le32 tv_sec;
	 *         le32 tv_nsec;
	 *     } mtime;
	 */

If you want to clean up and formalise all of these sorts of things, you might
need to invest in an rpcgen-like tool.  I've occasionally toyed with the idea
for afs in the kernel (I've hand-written all the marshalling/unmarshalling
code in fs/afs/fsclient.c, fs/afs/yfsclient.c and fs/afs/vlclient.c, but there
are some not-so-simple RPC calls to handle - FetchData and StoreData for
example).

David