Re: [PATCH] KVM: release anon file in failure path of vm creation

[Liu Shuo]

On Fri 15.Jul'16 at  3:26:03 +0100, Al Viro wrote:
> On Fri, Jul 15, 2016 at 10:22:04AM +0800, Liu Shuo wrote:
> > > You have no warranty whatsoever that descriptor table has not been changed
> > > by that point.  You should *NEVER* use sys_close() on failure exit paths
> > Could you please elaborate why we're not sure descriptor table's changing at the point?
>
> Because that could be called by one thread while another (having guessed the
> descriptor you are about to get) does close()/dup2()/etc.
If there is no such thread (who operates the descriptor based on
guessing), i can think the changing is safe at the point. As the fd has
not been delivered to userspace. Am i right?