From patchwork Sun Mar 27 03:44:04 2011 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: =?utf-8?q?Linus_L=C3=BCssing?= X-Patchwork-Id: 88491 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 ABEA2B6FB7 for ; Sun, 27 Mar 2011 14:44:29 +1100 (EST) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752680Ab1C0DoM (ORCPT ); Sat, 26 Mar 2011 23:44:12 -0400 Received: from fmmailgate03.web.de ([217.72.192.234]:52145 "EHLO fmmailgate03.web.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752645Ab1C0DoL (ORCPT ); Sat, 26 Mar 2011 23:44:11 -0400 Received: from smtp01.web.de ( [172.20.0.243]) by fmmailgate03.web.de (Postfix) with ESMTP id 0D5D718A7AFE5; Sun, 27 Mar 2011 05:44:09 +0200 (CEST) Received: from [94.134.27.247] (helo=localhost) by smtp01.web.de with asmtp (TLSv1:AES128-SHA:128) (WEB.DE 4.110 #2) id 1Q3gtc-0008Bl-00; Sun, 27 Mar 2011 05:44:08 +0200 Date: Sun, 27 Mar 2011 05:44:04 +0200 From: Linus =?utf-8?Q?L=C3=BCssing?= To: bridge@lists.linux-foundation.org Cc: Stephen Hemminger , David Miller , YOSHIFUJI Hideaki , Herbert Xu , netdev@vger.kernel.org Subject: Checksumming bug in bridge multicast snooping for IPv6? Message-ID: <20110327034404.GC31916@Sellars> MIME-Version: 1.0 Content-Disposition: inline User-Agent: Mutt/1.5.20 (2009-06-14) X-Sender: linus.luessing@web.de X-Provags-ID: V01U2FsdGVkX1+QjmhCyvy+8BepGN9Xjmx9+/tmjJzsCKR7JPZ9 b/Xu3biJxWHG9i4H9v//FCc9KDvj2snGNwUat16lSM5KFEPBKD LoSsNDwjdJIzM+F8wgAA== Sender: netdev-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: netdev@vger.kernel.org Hi everyone, Somehow I'm having trouble with the IPv6 bridge snooping again: MLDv2 Reports are dropped by the multicast snooping feature, looks like it has something to do with checksums. Wireshark does not display any weirdness, it at least reports the MLD reports checksum as correct. The setup is the following: The VM is running a current Linux version of torvalds branch with no other additions then the printk-debug patch attached (2.6.38+ #4 SMP PREEMPT Sat Mar 26 22:59:11 GMT 2011 i686 GNU/Linux): http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=tree;h=16c29dafcc86024048f1dbb8349d31cb22c7c55a;hb=16c29dafcc86024048f1dbb8349d31cb22c7c55a The host machine which is joining a multicast group is doing an explicit join on the KVM instances provided tap interface: IPv6: vlc -vvv "udp://@[ff12::124%vmtap1]" (IPv4: vlc -vvv "udp://@224.0.1.123") The host machine is running a kernel from Debian unstable: 2.6.37-2-amd64 #1 SMP Sun Feb 27 12:32:01 UTC 2011 x86_64 GNU/Linux See the attached debug patch and the according output for some more details where it fails (the bridge is basically ignoring the MLDv2 report due to the goto in this line: http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blob;f=net/bridge/br_multicast.c;h=f61eb2eff3fdd387b83d9fab642bb610dde1ad69;hb=HEAD#l1530) I'm also attaching both a wireshark capture of the ignored IPv6 MLDv2 report and the working IGMPv3 report, which correspond directly to the attached printk debug output. I'm a little bit startled because I definitely had that part working a couple of weeks ago and I'm still trying to figure out what I might have changed in the setup. I definitely have updated the VMs kernel, but the same issue is present for the 2.6.38 and also 2.6.37 release versions with my fixes backported (the latter one was the one I had been using back then). I probably have updated the multicast listener host's kernel, too, I might have been running 2.6.32 or something more earlier... I've also tried having the listener host in another VM with the same 2.6.38+ kernel as the bridge-snooping host, but also that did not make a difference. Anyways, skb_checksum_complete() is calculating the checksum from skb's data to tail pointer, right? RFC3810 for MLDv2, section 5.1.2 says: "The standard ICMPv6 checksum; it covers the entire MLDv2 message, plus a "pseudo-header" of IPv6 header fields [RFC2463]." Could it be that this "pseudo-header" is not included in the checksumming? Is there a function in the kernel which could already provide that? I guess that could also explain why it's working fine for IPv4, there it's just the IGMP message being checksummed according to RFC 3376, section 4.1.2. Cheers, Linus PS: There also seems to be another offset bug in the same function, see comment in debug patch file, though seemingly unrelated to the issue described above. Correcting that len-variable does to help for the above issue. IPv6: [ 2460.557303] +++ br_multicast_ipv6_rcv() [ 2460.558114] +++ br_multicast_ipv6_rcv() here 0.5 [ 2460.558114] +++ br_multicast_ipv6_rcv() len: 36, offset: 48, skb_network_offset(skb2): 0 [ 2460.558114] +++ br_multicast_ipv6_rcv() new len: -12 [ 2460.558114] +++ br_multicast_ipv6_rcv() skb2->len: 28 len: -12 skb->len: 76 [ 2460.558114] +++ br_multicast_ipv6_rcv() here 1.5 [ 2460.558114] +++ br_multicast_ipv6_rcv() here 1.7, skb2->csum is 0 [ 2460.558114] +++ br_multicast_ipv6_rcv() here 1.8, skb_checksum_complete(skb2): 81e9 (skb): 2e60 [ 2464.981808] +++ br_multicast_ipv6_rcv() [ 2464.982388] +++ br_multicast_ipv6_rcv() here 0.5 [ 2465.062634] +++ br_multicast_ipv6_rcv() len: 36, offset: 48, skb_network_offset(skb2): 0 [ 2465.066698] +++ br_multicast_ipv6_rcv() new len: -12 [ 2465.069183] +++ br_multicast_ipv6_rcv() skb2->len: 28 len: -12 skb->len: 76 [ 2465.072013] +++ br_multicast_ipv6_rcv() here 1.5 [ 2465.074217] +++ br_multicast_ipv6_rcv() here 1.7, skb2->csum is 0 [ 2465.076785] +++ br_multicast_ipv6_rcv() here 1.8, skb_checksum_complete(skb2): 81e9 (skb): 2e60 IPv4: [ 2325.265830] +++ br_multicast_ipv4_rcv() skb2->len: 40 len: 40 skb->len: 40 [ 2325.266567] +++ br_multicast_ipv4_rcv() 2) skb2->len: 16 len: 16 skb->len: 16 [ 2325.266567] +++ br_multicast_ipv4_rcv() here 1.7, skb2->csum is 0 [ 2325.266567] +++ br_multicast_ipv4_rcv() here 1.8, skb_checksum_complete: 0 [ 2325.266567] +++ br_ip4_multicast_add_group() eth1 224.0.1.123 [ 2327.326179] +++ br_multicast_ipv4_rcv() skb2->len: 40 len: 40 skb->len: 40 [ 2327.326674] +++ br_multicast_ipv4_rcv() 2) skb2->len: 16 len: 16 skb->len: 16 [ 2327.326674] +++ br_multicast_ipv4_rcv() here 1.7, skb2->csum is 0 [ 2327.326674] +++ br_multicast_ipv4_rcv() here 1.8, skb_checksum_complete: 0 [ 2327.326674] +++ br_ip4_multicast_add_group() eth1 224.0.1.123 diff --git a/net/bridge/br_multicast.c b/net/bridge/br_multicast.c index f61eb2e..3d4c5d2 100644 --- a/net/bridge/br_multicast.c +++ b/net/bridge/br_multicast.c @@ -764,8 +764,11 @@ static int br_ip4_multicast_add_group(struct net_bridge *br, { struct br_ip br_group; - if (ipv4_is_local_multicast(group)) +printk("+++ br_ip4_multicast_add_group() %s %pI4\n", port->dev->name, &group); + if (ipv4_is_local_multicast(group)) { +printk("+++ br_ip4_multicast_add_group() %s %pI4, is link local\n", port->dev->name, &group); return 0; + } br_group.u.ip4 = group; br_group.proto = htons(ETH_P_IP); @@ -780,8 +783,11 @@ static int br_ip6_multicast_add_group(struct net_bridge *br, { struct br_ip br_group; - if (!ipv6_is_transient_multicast(group)) +printk("+++ br_ip6_multicast_add_group() %s %pI6\n", port->dev->name, group); + if (!ipv6_is_transient_multicast(group)) { +printk("+++ br_ip6_multicast_add_group() %s %pI6, is not transient\n", port->dev->name, group); return 0; + } ipv6_addr_copy(&br_group.u.ip6, group); br_group.proto = htons(ETH_P_IPV6); @@ -1001,6 +1007,7 @@ static int br_ip6_multicast_mld2_report(struct net_bridge *br, int num; int err = 0; +printk("+++ br_ip6_multicast_mld2_report()\n"); if (!pskb_may_pull(skb, sizeof(*icmp6h))) return -EINVAL; @@ -1386,11 +1393,14 @@ static int br_multicast_ipv4_rcv(struct net_bridge *br, if (skb->len < len || len < ip_hdrlen(skb)) return -EINVAL; +printk("+++ br_multicast_ipv4_rcv() skb2->len: %i len: %i skb->len: %i\n", skb2->len, len, skb->len); if (skb->len > len) { +printk("+++ br_multicast_ipv4_rcv() doing clone\n"); skb2 = skb_clone(skb, GFP_ATOMIC); if (!skb2) return -ENOMEM; +printk("+++ br_multicast_ipv4_rcv() and pskb_trim_rcsum\n"); err = pskb_trim_rcsum(skb2, len); if (err) goto err_out; @@ -1405,14 +1415,20 @@ static int br_multicast_ipv4_rcv(struct net_bridge *br, if (!pskb_may_pull(skb2, sizeof(*ih))) goto out; +printk("+++ br_multicast_ipv4_rcv() 2) skb2->len: %i len: %i skb->len: %i\n", skb2->len, len, skb->len); switch (skb2->ip_summed) { case CHECKSUM_COMPLETE: +printk("+++ br_multicast_ipv4_rcv() here 1.6\n"); if (!csum_fold(skb2->csum)) break; /* fall through */ case CHECKSUM_NONE: +printk("+++ br_multicast_ipv4_rcv() here 1.7, skb2->csum is %x\n", skb2->csum); + __sum16 foobar; skb2->csum = 0; - if (skb_checksum_complete(skb2)) + foobar = skb_checksum_complete(skb2); +printk("+++ br_multicast_ipv4_rcv() here 1.8, skb_checksum_complete: %x\n", foobar); + if (foobar) goto out; } @@ -1459,6 +1475,7 @@ static int br_multicast_ipv6_rcv(struct net_bridge *br, int offset; int err; +printk("+++ br_multicast_ipv6_rcv()\n"); if (!pskb_may_pull(skb, sizeof(*ip6h))) return -EINVAL; @@ -1476,6 +1493,7 @@ static int br_multicast_ipv6_rcv(struct net_bridge *br, return 0; len = ntohs(ip6h->payload_len); +// len = ntohs(ip6h->payload_len) + sizeof(*ip6h); <- should probably be this? if (skb->len < len) return -EINVAL; @@ -1485,6 +1503,7 @@ static int br_multicast_ipv6_rcv(struct net_bridge *br, if (offset < 0 || nexthdr != IPPROTO_ICMPV6) return 0; +printk("+++ br_multicast_ipv6_rcv() here 0.5\n"); /* Okay, we found ICMPv6 header */ skb2 = skb_clone(skb, GFP_ATOMIC); if (!skb2) @@ -1494,7 +1513,9 @@ static int br_multicast_ipv6_rcv(struct net_bridge *br, if (!pskb_may_pull(skb2, offset + sizeof(struct icmp6hdr))) goto out; +printk("+++ br_multicast_ipv6_rcv() len: %i, offset: %i, skb_network_offset(skb2): %i\n", len, offset, skb_network_offset(skb2)); len -= offset - skb_network_offset(skb2); +printk("+++ br_multicast_ipv6_rcv() new len: %i\n", len); __skb_pull(skb2, offset); skb_reset_transport_header(skb2); @@ -1513,27 +1534,37 @@ static int br_multicast_ipv6_rcv(struct net_bridge *br, } /* Okay, we found MLD message. Check further. */ +printk("+++ br_multicast_ipv6_rcv() skb2->len: %i len: %i skb->len: %i\n", skb2->len, len, skb->len); if (skb2->len > len) { +printk("+++ br_multicast_ipv6_rcv() doing pskb_trim_rcsum\n"); err = pskb_trim_rcsum(skb2, len); if (err) goto out; } +printk("+++ br_multicast_ipv6_rcv() here 1.5\n"); switch (skb2->ip_summed) { case CHECKSUM_COMPLETE: +printk("+++ br_multicast_ipv6_rcv() here 1.6\n"); if (!csum_fold(skb2->csum)) break; /*FALLTHROUGH*/ case CHECKSUM_NONE: +printk("+++ br_multicast_ipv6_rcv() here 1.7, skb2->csum is %x\n", skb2->csum); + __sum16 foobar; skb2->csum = 0; - if (skb_checksum_complete(skb2)) + foobar = skb_checksum_complete(skb2); + if (foobar) { +printk("+++ br_multicast_ipv6_rcv() here 1.8, skb_checksum_complete(skb2): %x (skb): %x\n", foobar, skb_checksum_complete(skb)); goto out; + } } err = 0; BR_INPUT_SKB_CB(skb)->igmp = 1; +printk("+++ br_multicast_ipv6_rcv() here 2\n"); switch (icmp6h->icmp6_type) { case ICMPV6_MGM_REPORT: {