Quoting Serge E. Hallyn (se...@us.ibm.com):
> Quoting Dan Smith (da...@us.ibm.com):
> > SH> ckpt_hdr_put(ctx, in) ?
> >
> > Oops.
> >
> > >> + un->this = ckpt_obj_lookup_add(ctx, sk, CKPT_OBJ_SOCK, &new);
> > >> + if (un->this < 0)
> > >> + goto out;
> > >> +
> > >> + if (sk->peer)
> > >> + un->peer = ckpt_obj_lookup_add(ctx, pr, CKPT_OBJ_SOCK,
> > >> &new);
> > >> + else
> > >> + un->peer = 0;
> > >> +
> > >> + if (un->peer < 0) {
> > >> + ret = un->peer;
> > >> + goto out;
> > >> + }
> >
> > SH> So what if new == 1 for either un->this or un->peer? You never
> > SH> actually write them out to the checkpoint image?
> >
> > On the checkpoint run, the new flag doesn't matter to us (and isn't
> > used here). Am I missing something?
>
> Well when is do_sock_file_checkpoint() going to be called?
Ok, I see, h is the actual socket details represented by
un->this. I was glossing over the writing of h thinking
it was just a ckpt_hdr.
Looks good.
thanks,
-serge
_______________________________________________
Containers mailing list
contain...@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containers
_______________________________________________
Devel mailing list
Devel@openvz.org
https://openvz.org/mailman/listinfo/devel
