]> git.proxmox.com Git - mirror_ubuntu-zesty-kernel.git/commitdiff
ipv6: Don't increase IPSTATS_MIB_FRAGFAILS twice in ip6_fragment()
authorStefano Brivio <sbrivio@redhat.com>
Mon, 24 Jul 2017 21:14:28 +0000 (23:14 +0200)
committerThadeu Lima de Souza Cascardo <cascardo@canonical.com>
Tue, 22 Aug 2017 09:58:14 +0000 (06:58 -0300)
BugLink: http://bugs.launchpad.net/bugs/1711526
[ Upstream commit afce615aaabfbaad02550e75c0bec106dafa1adf ]

RFC 2465 defines ipv6IfStatsOutFragFails as:

"The number of IPv6 datagrams that have been discarded
 because they needed to be fragmented at this output
 interface but could not be."

The existing implementation, instead, would increase the counter
twice in case we fail to allocate room for single fragments:
once for the fragment, once for the datagram.

This didn't look intentional though. In one of the two affected
affected failure paths, the double increase was simply a result
of a new 'goto fail' statement, introduced to avoid a skb leak.
The other path appears to be affected since at least 2.6.12-rc2.

Reported-by: Sabrina Dubroca <sdubroca@redhat.com>
Fixes: 1d325d217c7f ("ipv6: ip6_fragment: fix headroom tests and skb leak")
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@canonical.com>
net/ipv6/ip6_output.c

index 4a4840cda6b8c6ed5cf249c6737dad73e18a1bc2..5f58e058529ea53ea46df686eacff38fde245df6 100644 (file)
@@ -647,8 +647,6 @@ int ip6_fragment(struct net *net, struct sock *sk, struct sk_buff *skb,
                *prevhdr = NEXTHDR_FRAGMENT;
                tmp_hdr = kmemdup(skb_network_header(skb), hlen, GFP_ATOMIC);
                if (!tmp_hdr) {
-                       IP6_INC_STATS(net, ip6_dst_idev(skb_dst(skb)),
-                                     IPSTATS_MIB_FRAGFAILS);
                        err = -ENOMEM;
                        goto fail;
                }
@@ -767,8 +765,6 @@ slow_path:
                frag = alloc_skb(len + hlen + sizeof(struct frag_hdr) +
                                 hroom + troom, GFP_ATOMIC);
                if (!frag) {
-                       IP6_INC_STATS(net, ip6_dst_idev(skb_dst(skb)),
-                                     IPSTATS_MIB_FRAGFAILS);
                        err = -ENOMEM;
                        goto fail;
                }