From patchwork Sat Sep 21 04:27:00 2013 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Hannes Frederic Sowa X-Patchwork-Id: 276835 X-Patchwork-Delegate: davem@davemloft.net Return-Path: X-Original-To: patchwork-incoming@ozlabs.org Delivered-To: patchwork-incoming@ozlabs.org Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by ozlabs.org (Postfix) with ESMTP id 611322C01B2 for ; Sat, 21 Sep 2013 14:27:31 +1000 (EST) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752090Ab3IUE1N (ORCPT ); Sat, 21 Sep 2013 00:27:13 -0400 Received: from order.stressinduktion.org ([87.106.68.36]:41523 "EHLO order.stressinduktion.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751692Ab3IUE1M (ORCPT ); Sat, 21 Sep 2013 00:27:12 -0400 Received: by order.stressinduktion.org (Postfix, from userid 500) id 7EC2D1A0C29D; Sat, 21 Sep 2013 06:27:01 +0200 (CEST) Date: Sat, 21 Sep 2013 06:27:00 +0200 From: Hannes Frederic Sowa To: netdev@vger.kernel.org Cc: yoshfuji@linux-ipv6.org, davem@davemloft.net Subject: [PATCH] ipv6: udp packets following an UFO enqueued packet need also be handled by UFO Message-ID: <20130921042700.GB8070@order.stressinduktion.org> Mail-Followup-To: netdev@vger.kernel.org, yoshfuji@linux-ipv6.org, davem@davemloft.net Mime-Version: 1.0 Content-Disposition: inline Sender: netdev-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: netdev@vger.kernel.org In the following scenario the socket is corked: If the first UDP packet is larger then the mtu we try to append it to the write queue via ip6_ufo_append_data. A following packet, which is smaller than the mtu would be appended to the already queued up gso-skb via plain ip6_append_data. This causes random memory corruptions. In ip6_ufo_append_data we also have to be careful to not queue up the same skb multiple times. So setup the gso frame only when no first skb is available. This also fixes a shortcoming where we add the current packet's length to cork->length but return early because of a packet > mtu with dontfrag set (instead of sutracting it again). Found with trinity. Cc: YOSHIFUJI Hideaki Signed-off-by: Hannes Frederic Sowa Reported-by: Dmitry Vyukov --- I could only test this with virtualized UFO enabled network cards. Could someone test this on real hardware? net/ipv6/ip6_output.c | 53 +++++++++++++++++++++------------------------------ 1 file changed, 22 insertions(+), 31 deletions(-) diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c index 3a692d5..a54c45c 100644 --- a/net/ipv6/ip6_output.c +++ b/net/ipv6/ip6_output.c @@ -1015,6 +1015,8 @@ static inline int ip6_ufo_append_data(struct sock *sk, * udp datagram */ if ((skb = skb_peek_tail(&sk->sk_write_queue)) == NULL) { + struct frag_hdr fhdr; + skb = sock_alloc_send_skb(sk, hh_len + fragheaderlen + transhdrlen + 20, (flags & MSG_DONTWAIT), &err); @@ -1036,12 +1038,6 @@ static inline int ip6_ufo_append_data(struct sock *sk, skb->protocol = htons(ETH_P_IPV6); skb->ip_summed = CHECKSUM_PARTIAL; skb->csum = 0; - } - - err = skb_append_datato_frags(sk,skb, getfrag, from, - (length - transhdrlen)); - if (!err) { - struct frag_hdr fhdr; /* Specify the length of each IPv6 datagram fragment. * It has to be a multiple of 8. @@ -1052,15 +1048,10 @@ static inline int ip6_ufo_append_data(struct sock *sk, ipv6_select_ident(&fhdr, rt); skb_shinfo(skb)->ip6_frag_id = fhdr.identification; __skb_queue_tail(&sk->sk_write_queue, skb); - - return 0; } - /* There is not enough support do UPD LSO, - * so follow normal path - */ - kfree_skb(skb); - return err; + return skb_append_datato_frags(sk, skb, getfrag, from, + (length - transhdrlen)); } static inline struct ipv6_opt_hdr *ip6_opt_dup(struct ipv6_opt_hdr *src, @@ -1227,27 +1218,27 @@ int ip6_append_data(struct sock *sk, int getfrag(void *from, char *to, * --yoshfuji */ - cork->length += length; - if (length > mtu) { - int proto = sk->sk_protocol; - if (dontfrag && (proto == IPPROTO_UDP || proto == IPPROTO_RAW)){ - ipv6_local_rxpmtu(sk, fl6, mtu-exthdrlen); - return -EMSGSIZE; - } - - if (proto == IPPROTO_UDP && - (rt->dst.dev->features & NETIF_F_UFO)) { + if ((length > mtu) && dontfrag && (sk->sk_protocol == IPPROTO_UDP || + sk->sk_protocol == IPPROTO_RAW)) { + ipv6_local_rxpmtu(sk, fl6, mtu-exthdrlen); + return -EMSGSIZE; + } - err = ip6_ufo_append_data(sk, getfrag, from, length, - hh_len, fragheaderlen, - transhdrlen, mtu, flags, rt); - if (err) - goto error; - return 0; - } + skb = skb_peek_tail(&sk->sk_write_queue); + cork->length += length; + if (((length > mtu) || + (skb && skb_is_gso(skb))) && + (sk->sk_protocol == IPPROTO_UDP) && + (rt->dst.dev->features & NETIF_F_UFO)) { + err = ip6_ufo_append_data(sk, getfrag, from, length, + hh_len, fragheaderlen, + transhdrlen, mtu, flags, rt); + if (err) + goto error; + return 0; } - if ((skb = skb_peek_tail(&sk->sk_write_queue)) == NULL) + if (!skb) goto alloc_new_skb; while (length > 0) {