Re: [PATCH 1/3] xen/blkback: Support 'feature-barrier' aka old-style BARRIER requests.

[Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>]

On Mon, Oct 17, 2011 at 12:36:18PM +0100, Jan Beulich wrote:
> >>> On 10.10.11 at 17:28, Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> wrote:
> > @@ -481,6 +503,10 @@ static void __end_block_io_op(struct pending_req 
> > *pending_req, int error)
> >  			      pending_req->operation, pending_req->status);
> >  		xen_blkif_put(pending_req->blkif);
> >  		free_req(pending_req);
> > +		if (atomic_read(&pending_req->blkif->refcnt) <= 2) {
> > +			if (atomic_read(&pending_req->blkif->drain))
> > +				complete(&pending_req->blkif->drain_complete);
> > +		}
> 
> Shouldn't this be done *before* the call the free_req()? Or

Yes, otherwise we could referencing somebody's else blkif->refcnt.

Thanks for spotting that.
> alternatively a local copy of pending_req->blkif be obtained before
> freeing pending_req (which could be used in a couple more places
> in this function)?
> 
> Jan