On Mon, Apr 09, 2018 at 11:00:06AM -0700, Chenbo Feng wrote:
>On Sun, Apr 8, 2018 at 5:28 PM, Sasha Levin
><Alexander.Levin@xxxxxxxxxxxxx> wrote:
>> From: Chenbo Feng <fengc@xxxxxxxxxx>
>>
>> [ Upstream commit 8fac365f63c866a00015fa13932d8ffc584518b8 ]
>>
>> Currently in both ipv4 and ipv6 code path, the ack packet received when
>> sk at TCP_NEW_SYN_RECV state is not filtered by socket filter or cgroup
>> filter since it is handled from tcp_child_process and never reaches the
>> tcp_filter inside tcp_v4_rcv or tcp_v6_rcv. Adding a tcp_filter hooks
>> here can make sure all the ingress tcp packet can be correctly filtered.
>>
>> Signed-off-by: Chenbo Feng <fengc@xxxxxxxxxx>
>> Signed-off-by: David S. Miller <davem@xxxxxxxxxxxxx>
>> Signed-off-by: Sasha Levin <alexander.levin@xxxxxxxxxxxxx>
>> ---
>> net/ipv4/tcp_ipv4.c | 2 ++
>> net/ipv6/tcp_ipv6.c | 2 ++
>> 2 files changed, 4 insertions(+)
>>
>> diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
>> index 61c93a93f228..2eb9a1747f25 100644
>> --- a/net/ipv4/tcp_ipv4.c
>> +++ b/net/ipv4/tcp_ipv4.c
>> @@ -1639,6 +1639,8 @@ process:
>> }
>> if (nsk == sk) {
>> reqsk_put(req);
>> + } else if (tcp_filter(sk, skb)) {
>> + goto discard_and_relse;
>> } else if (tcp_child_process(sk, nsk, skb)) {
>> tcp_v4_send_reset(nsk, skb);
>> goto discard_and_relse;
>> diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
>> index 74cbcc4b399c..7d7f0999a919 100644
>> --- a/net/ipv6/tcp_ipv6.c
>> +++ b/net/ipv6/tcp_ipv6.c
>> @@ -1428,6 +1428,8 @@ process:
>> if (nsk == sk) {
>> reqsk_put(req);
>> tcp_v6_restore_cb(skb);
>> + } else if (tcp_filter(sk, skb)) {
>> + goto discard_and_relse;
>> } else if (tcp_child_process(sk, nsk, skb)) {
>> tcp_v6_send_reset(nsk, skb);
>> goto discard_and_relse;
>> --
>> 2.15.1
>There is a bug fix for this patch upstream:
>"d624d276d1ddacbcb12ad96832ce0c7b82cd25db tcp: fix possible deadlock
>in TCP stack vs BPF filter", Please make you include that as well if
>you want to add this patch to stable.
That commit had quite a few conflicts, so I dropped this commit instead.
Thanks!