udp: properly support MSG_PEEK with truncated buffers
authorEric Dumazet <edumazet@google.com>
Wed, 30 Dec 2015 13:51:12 +0000 (08:51 -0500)
committerManish Tuteja <mtuteja@nvidia.com>
Mon, 17 Apr 2017 23:55:59 +0000 (16:55 -0700)
commit4e811459502fbf165d5148f9b0f2cdae03077a45
tree33d7126500df89b5251fd473b97783a213deb8e2
parentf3d53da51f9f44437285ca91863b7e5a6e9f08db
udp: properly support MSG_PEEK with truncated buffers

Backport of this upstream commit into stable kernels :
89c22d8c3b27 ("net: Fix skb csum races when peeking")
exposed a bug in udp stack vs MSG_PEEK support, when user provides
a buffer smaller than skb payload.

In this case,
skb_copy_and_csum_datagram_iovec(skb, sizeof(struct udphdr),
                                 msg->msg_iov);
returns -EFAULT.

This bug does not happen in upstream kernels since Al Viro did a great
job to replace this into :
skb_copy_and_csum_datagram_msg(skb, sizeof(struct udphdr), msg);
This variant is safe vs short buffers.

For the time being, instead reverting Herbert Xu patch and add back
skb->ip_summed invalid changes, simply store the result of
udp_lib_checksum_complete() so that we avoid computing the checksum a
second time, and avoid the problematic
skb_copy_and_csum_datagram_iovec() call.

This patch can be applied on recent kernels as it avoids a double
checksumming, then backported to stable kernels as a bug fix.

Bug 1885879

Change-Id: Ie7215bf99beee2fccd662152e80767cdeb6ff9b2
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Gagan Grover <ggrover@nvidia.com>
Reviewed-on: http://git-master/r/1330675
Reviewed-by: Vinayak Pane <vpane@nvidia.com>
GVS: Gerrit_Virtual_Submit
net/ipv4/udp.c
net/ipv6/udp.c