net_sched: unset TCQ_F_CAN_BYPASS when adding filters
authorCong Wang <xiyou.wangcong@gmail.com>
Tue, 16 Jul 2019 20:57:30 +0000 (13:57 -0700)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Sun, 28 Jul 2019 06:29:27 +0000 (08:29 +0200)
commitd9571a9f5ec19820571de3c2244d16f5b53b8bf0
tree9c7158ce60e2b027146c97ab2dc68ebf07eae1d1
parent69cd584546159c82e869cef15675cfe191a649e1
net_sched: unset TCQ_F_CAN_BYPASS when adding filters

[ Upstream commit 3f05e6886a595c9a29a309c52f45326be917823c ]

For qdisc's that support TC filters and set TCQ_F_CAN_BYPASS,
notably fq_codel, it makes no sense to let packets bypass the TC
filters we setup in any scenario, otherwise our packets steering
policy could not be enforced.

This can be reproduced easily with the following script:

 ip li add dev dummy0 type dummy
 ifconfig dummy0 up
 tc qd add dev dummy0 root fq_codel
 tc filter add dev dummy0 parent 8001: protocol arp basic action mirred egress redirect dev lo
 tc filter add dev dummy0 parent 8001: protocol ip basic action mirred egress redirect dev lo
 ping -I dummy0 192.168.112.1

Without this patch, packets are sent directly to dummy0 without
hitting any of the filters. With this patch, packets are redirected
to loopback as expected.

This fix is not perfect, it only unsets the flag but does not set it back
because we have to save the information somewhere in the qdisc if we
really want that. Note, both fq_codel and sfq clear this flag in their
->bind_tcf() but this is clearly not sufficient when we don't use any
class ID.

Fixes: 23624935e0c4 ("net_sched: TCQ_F_CAN_BYPASS generalization")
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
net/sched/cls_api.c
net/sched/sch_fq_codel.c
net/sched/sch_sfq.c