Thank you for the review!
On Thu, Aug 08, 2013 at 05:01:46PM -0700, Ansis Atteka wrote:
> git am
> ~/Downloads/ovs-dev-netdev-22-27-netdev-linux-netdev-bsd-Make-access-to-AF_INET-socket-thread-safe..patch
> Applying: netdev-linux, netdev-bsd: Make access to AF_INET socket
> thread-safe. /home/aatteka/git/openvswitch/.git/rebase-apply/patch:639: new
> blank line at EOF. +
Hmm. I could not find this in my patch. I have rebased a few times
since I posted the patch, so maybe I have already fixed this.
> On Thu, Aug 1, 2013 at 2:29 PM, Ben Pfaff <b...@nicira.com> wrote:
> > +/* Calls ioctl() on an AF_INET sock, passing the specified 'command' and
> > + * 'arg'. Returns 0 if successful, otherwise a positive errno value. */
> > +int
> > +af_inet_ioctl(int command, const void *arg)
> > +{
> > + static struct ovsthread_once once = OVSTHREAD_ONCE_INITIALIZER;
> > + static int sock;
> > +
> > + if (ovsthread_once_start(&once)) {
> >
> If socket() failed, then do we want to retry it on the next af_inet_ioctl()
> call?
>
> I guess that might happen, for example, if we temporarily ran out of file
> descriptors.
I think that it is unlikely enough that it is not worth worrying about
(famous last words, I know).
> > +int
> > +af_inet_ifreq_ioctl(const char *name, struct ifreq *ifr, int cmd,
> > + const char *cmd_name)
> > +{
> > + int error;
> > +
> > + ovs_strzcpy(ifr->ifr_name, name, sizeof ifr->ifr_name);
> > + error = af_inet_ioctl(cmd, ifr);
> > + if (error) {
> > + static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(5, 20);
> > + VLOG_DBG_RL(&rl, "%s: ioctl(%s) failed: %s", name, cmd_name,
> > + ovs_strerror(error));
> >
> it looks like there is one unnecessary space
Thank you. I fixed this.
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
