On 3. Jan 2012, at 22:22 , John Baldwin wrote:

> On Tuesday, January 03, 2012 5:14:22 pm Hiroki Sato wrote:
>> John Baldwin <j...@freebsd.org> wrote
>>  in <201201031608.59688....@freebsd.org>:
>> 
>> jh> > With this patch in_lifaddr_ioctl() now looks more syntactically similar
>> jh> > to in6_lifaddr_ioctl(). They could look even more similar by 
>> eliminating
>> jh> > a lot of whitespace changes present here or there.
>> jh>
>> jh> Hmmm.  Actually, it seems to be a bit more broken.  Note that it is 
>> expecting
>> jh> to get a sockaddr_in, but it is checking for AF_INET6, not AF_INET in its
>> jh> loop.  That bug seems to go back to the original import from KAME.  I'm 
>> not
>> jh> sure if the two can be merged since they work on different underyling 
>> data
>> jh> structures though.
>> 
>> Hmm, a fix for that bug was not merged for some reason.  Something
>> like the attached patch should be applied.
> 
> Ah, great, I've merged that into the patch, thanks!
> 
> Index: in.c
> ===================================================================
> --- in.c      (revision 229406)
> +++ in.c      (working copy)
> @@ -735,7 +735,7 @@ in_lifaddr_ioctl(struct socket *so, u_long cmd, ca
>               if (iflr->flags & IFLR_PREFIX)
>                       return (EINVAL);
> 
> -             /* copy args to in_aliasreq, perform ioctl(SIOCAIFADDR_IN6). */
> +             /* copy args to in_aliasreq, perform ioctl(SIOCAIFADDR). */
>               bzero(&ifra, sizeof(ifra));
>               bcopy(iflr->iflr_name, ifra.ifra_name,
>                       sizeof(ifra.ifra_name));
> @@ -784,8 +784,9 @@ in_lifaddr_ioctl(struct socket *so, u_long cmd, ca
>                       }
>               }
> 
> +             IF_ADDR_LOCK(ifp);
>               TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) {
> -                     if (ifa->ifa_addr->sa_family != AF_INET6)
> +                     if (ifa->ifa_addr->sa_family != AF_INET)
>                               continue;
>                       if (match.s_addr == 0)
>                               break;

When doing "noise changes" like fixing comment in the same go (I'd prefer
that to be independent of the locking/refcount changes if possible, you
could also wrap the long line:
    792                         candidate.s_addr = ((struct sockaddr_in 
*)&ifa->ifa_addr)->sin_addr.s_addr;


Otherwise looks good.


> @@ -794,6 +795,9 @@ in_lifaddr_ioctl(struct socket *so, u_long cmd, ca
>                       if (candidate.s_addr == match.s_addr)
>                               break;
>               }
> +             if (ifa != NULL)
> +                     ifa_ref(ifa);
> +             IF_ADDR_UNLOCK(ifp);
>               if (ifa == NULL)
>                       return (EADDRNOTAVAIL);
>               ia = (struct in_ifaddr *)ifa;
> @@ -812,12 +816,13 @@ in_lifaddr_ioctl(struct socket *so, u_long cmd, ca
>                               in_mask2len(&ia->ia_sockmask.sin_addr);
> 
>                       iflr->flags = 0;        /*XXX*/
> +                     ifa_free(ifa);
> 
>                       return (0);
>               } else {
>                       struct in_aliasreq ifra;
> 
> -                     /* fill in_aliasreq and do ioctl(SIOCDIFADDR_IN6) */
> +                     /* fill in_aliasreq and do ioctl(SIOCDIFADDR) */
>                       bzero(&ifra, sizeof(ifra));
>                       bcopy(iflr->iflr_name, ifra.ifra_name,
>                               sizeof(ifra.ifra_name));
> @@ -830,6 +835,7 @@ in_lifaddr_ioctl(struct socket *so, u_long cmd, ca
>                       }
>                       bcopy(&ia->ia_sockmask, &ifra.ifra_dstaddr,
>                               ia->ia_sockmask.sin_len);
> +                     ifa_free(ifa);
> 
>                       return (in_control(so, SIOCDIFADDR, (caddr_t)&ifra,
>                           ifp, td));
> 
> -- 
> John Baldwin
> _______________________________________________
> freebsd-net@freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-net
> To unsubscribe, send any mail to "freebsd-net-unsubscr...@freebsd.org"

-- 
Bjoern A. Zeeb                                 You have to have visions!
   It does not matter how good you are. It matters what good you do!

_______________________________________________
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to "freebsd-net-unsubscr...@freebsd.org"

Reply via email to