Date: Tue, 8 Jul 2014 21:24:06 -0700 From: Navdeep Parhar <navdeep@chelsio.com> To: Gleb Smirnoff <glebius@FreeBSD.org>, Alan Somers <asomers@FreeBSD.org> Cc: "freebsd-net@freebsd.org" <freebsd-net@FreeBSD.org> Subject: Re: ifaddr refcount problem Message-ID: <20140709042406.GA64350@dwarf.asicdesigners.com> In-Reply-To: <20140625100835.GK28199@glebius.int.ru> References: <53A48849.8080504@chelsio.com> <20140623085229.GQ28199@FreeBSD.org> <CAOtMX2hGpvhK5TkBUfXvXisu5E2zen=h6MEv_Bxecn=aKsnnnQ@mail.gmail.com> <20140624090847.GB28199@glebius.int.ru> <CAOtMX2gV%2BnTLs-%2BpFWWETLKW2Dtrtxb-Z30Dq65Fctwy4pBfPA@mail.gmail.com> <20140625100835.GK28199@glebius.int.ru>
next in thread | previous in thread | raw e-mail | index | archive | help
I got distracted by some other issues and lost track of this thread.
Can one of you please commit the fix if the discussion has reached a
conclusion? Thanks!
Regards,
Navdeep
On Wed, Jun 25, 2014 at 02:08:35PM +0400, Gleb Smirnoff wrote:
> Alan,
>
> On Tue, Jun 24, 2014 at 08:43:40AM -0600, Alan Somers wrote:
> A> That looks better. But I think there is one more possibility for a
> A> leak. For multicast packets, IFP_TO_IA at line 263 will call
> A> ifa_ref(), unless the the interface has no address assigned. How
> A> about this patch instead? Also, not tested.
>
> Again you are right, thanks.
>
> The patch needs to reset have_ia_ref in the 'goto again' case.
>
> Also, I'd suggest a tiny change to your patch so that we don't
> have a broken logic reading the code as "I have ia reference,
> but don't have ia".
>
> --
> Totus tuus, Glebius.
> Index: ip_output.c
> ===================================================================
> --- ip_output.c (revision 267536)
> +++ ip_output.c (working copy)
> @@ -136,6 +136,7 @@ ip_output(struct mbuf *m, struct mbuf *opt, struct
> struct rtentry *rte; /* cache for ro->ro_rt */
> struct in_addr odst;
> struct m_tag *fwd_tag = NULL;
> + int have_ia_ref;
> #ifdef IPSEC
> int no_route_but_check_spd = 0;
> #endif
> @@ -202,6 +203,7 @@ ip_output(struct mbuf *m, struct mbuf *opt, struct
> gw = dst = (struct sockaddr_in *)&ro->ro_dst;
> again:
> ia = NULL;
> + have_ia_ref = 0;
> /*
> * If there is a cached route, check that it is to the same
> * destination and is still up. If not, free it and try again.
> @@ -238,6 +240,7 @@ again:
> error = ENETUNREACH;
> goto bad;
> }
> + have_ia_ref = 1;
> ip->ip_dst.s_addr = INADDR_BROADCAST;
> dst->sin_addr = ip->ip_dst;
> ifp = ia->ia_ifp;
> @@ -250,6 +253,7 @@ again:
> error = ENETUNREACH;
> goto bad;
> }
> + have_ia_ref = 1;
> ifp = ia->ia_ifp;
> ip->ip_ttl = 1;
> isbroadcast = in_broadcast(dst->sin_addr, ifp);
> @@ -261,6 +265,8 @@ again:
> */
> ifp = imo->imo_multicast_ifp;
> IFP_TO_IA(ifp, ia);
> + if (ia)
> + have_ia_ref = 1;
> isbroadcast = 0; /* fool gcc */
> } else {
> /*
> @@ -552,8 +558,11 @@ sendit:
> #endif
> error = netisr_queue(NETISR_IP, m);
> goto done;
> - } else
> + } else {
> + if (have_ia_ref)
> + ifa_free(&ia->ifa);
> goto again; /* Redo the routing table lookup. */
> + }
> }
>
> /* See if local, if yes, send it to netisr with IP_FASTFWD_OURS. */
> @@ -582,6 +591,8 @@ sendit:
> m->m_flags |= M_SKIP_FIREWALL;
> m->m_flags &= ~M_IP_NEXTHOP;
> m_tag_delete(m, fwd_tag);
> + if (have_ia_ref)
> + ifa_free(&ia->ifa);
> goto again;
> }
>
> @@ -694,6 +705,8 @@ passout:
> done:
> if (ro == &iproute)
> RO_RTFREE(ro);
> + if (have_ia_ref)
> + ifa_free(&ia->ifa);
> return (error);
> bad:
> m_freem(m);
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20140709042406.GA64350>