linux/net
Daniel Borkmann 89e5c58fc1 net: Fix gro aggregation for udp encaps with zero csum
We noticed a GRO issue for UDP-based encaps such as vxlan/geneve when the
csum for the UDP header itself is 0. In that case, GRO aggregation does
not take place on the phys dev, but instead is deferred to the vxlan/geneve
driver (see trace below).

The reason is essentially that GRO aggregation bails out in udp_gro_receive()
for such case when drivers marked the skb with CHECKSUM_UNNECESSARY (ice, i40e,
others) where for non-zero csums 2abb7cdc0d ("udp: Add support for doing
checksum unnecessary conversion") promotes those skbs to CHECKSUM_COMPLETE
and napi context has csum_valid set. This is however not the case for zero
UDP csum (here: csum_cnt is still 0 and csum_valid continues to be false).

At the same time 57c67ff4bd ("udp: additional GRO support") added matches
on !uh->check ^ !uh2->check as part to determine candidates for aggregation,
so it certainly is expected to handle zero csums in udp_gro_receive(). The
purpose of the check added via 662880f442 ("net: Allow GRO to use and set
levels of checksum unnecessary") seems to catch bad csum and stop aggregation
right away.

One way to fix aggregation in the zero case is to only perform the !csum_valid
check in udp_gro_receive() if uh->check is infact non-zero.

Before:

  [...]
  swapper     0 [008]   731.946506: net:netif_receive_skb: dev=enp10s0f0  skbaddr=0xffff966497100400 len=1500   (1)
  swapper     0 [008]   731.946507: net:netif_receive_skb: dev=enp10s0f0  skbaddr=0xffff966497100200 len=1500
  swapper     0 [008]   731.946507: net:netif_receive_skb: dev=enp10s0f0  skbaddr=0xffff966497101100 len=1500
  swapper     0 [008]   731.946508: net:netif_receive_skb: dev=enp10s0f0  skbaddr=0xffff966497101700 len=1500
  swapper     0 [008]   731.946508: net:netif_receive_skb: dev=enp10s0f0  skbaddr=0xffff966497101b00 len=1500
  swapper     0 [008]   731.946508: net:netif_receive_skb: dev=enp10s0f0  skbaddr=0xffff966497100600 len=1500
  swapper     0 [008]   731.946508: net:netif_receive_skb: dev=enp10s0f0  skbaddr=0xffff966497100f00 len=1500
  swapper     0 [008]   731.946509: net:netif_receive_skb: dev=enp10s0f0  skbaddr=0xffff966497100a00 len=1500
  swapper     0 [008]   731.946516: net:netif_receive_skb: dev=enp10s0f0  skbaddr=0xffff966497100500 len=1500
  swapper     0 [008]   731.946516: net:netif_receive_skb: dev=enp10s0f0  skbaddr=0xffff966497100700 len=1500
  swapper     0 [008]   731.946516: net:netif_receive_skb: dev=enp10s0f0  skbaddr=0xffff966497101d00 len=1500   (2)
  swapper     0 [008]   731.946517: net:netif_receive_skb: dev=enp10s0f0  skbaddr=0xffff966497101000 len=1500
  swapper     0 [008]   731.946517: net:netif_receive_skb: dev=enp10s0f0  skbaddr=0xffff966497101c00 len=1500
  swapper     0 [008]   731.946517: net:netif_receive_skb: dev=enp10s0f0  skbaddr=0xffff966497101400 len=1500
  swapper     0 [008]   731.946518: net:netif_receive_skb: dev=enp10s0f0  skbaddr=0xffff966497100e00 len=1500
  swapper     0 [008]   731.946518: net:netif_receive_skb: dev=enp10s0f0  skbaddr=0xffff966497101600 len=1500
  swapper     0 [008]   731.946521: net:netif_receive_skb: dev=enp10s0f0  skbaddr=0xffff966497100800 len=774
  swapper     0 [008]   731.946530: net:netif_receive_skb: dev=test_vxlan skbaddr=0xffff966497100400 len=14032 (1)
  swapper     0 [008]   731.946530: net:netif_receive_skb: dev=test_vxlan skbaddr=0xffff966497101d00 len=9112  (2)
  [...]

  # netperf -H 10.55.10.4 -t TCP_STREAM -l 20
  MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 10.55.10.4 () port 0 AF_INET : demo
  Recv   Send    Send
  Socket Socket  Message  Elapsed
  Size   Size    Size     Time     Throughput
  bytes  bytes   bytes    secs.    10^6bits/sec

   87380  16384  16384    20.01    13129.24

After:

  [...]
  swapper     0 [026]   521.862641: net:netif_receive_skb: dev=enp10s0f0  skbaddr=0xffff93ab0d479000 len=11286 (1)
  swapper     0 [026]   521.862643: net:netif_receive_skb: dev=test_vxlan skbaddr=0xffff93ab0d479000 len=11236 (1)
  swapper     0 [026]   521.862650: net:netif_receive_skb: dev=enp10s0f0  skbaddr=0xffff93ab0d478500 len=2898  (2)
  swapper     0 [026]   521.862650: net:netif_receive_skb: dev=enp10s0f0  skbaddr=0xffff93ab0d479f00 len=8490  (3)
  swapper     0 [026]   521.862653: net:netif_receive_skb: dev=test_vxlan skbaddr=0xffff93ab0d478500 len=2848  (2)
  swapper     0 [026]   521.862653: net:netif_receive_skb: dev=test_vxlan skbaddr=0xffff93ab0d479f00 len=8440  (3)
  [...]

  # netperf -H 10.55.10.4 -t TCP_STREAM -l 20
  MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 10.55.10.4 () port 0 AF_INET : demo
  Recv   Send    Send
  Socket Socket  Message  Elapsed
  Size   Size    Size     Time     Throughput
  bytes  bytes   bytes    secs.    10^6bits/sec

   87380  16384  16384    20.01    24576.53

Fixes: 57c67ff4bd ("udp: additional GRO support")
Fixes: 662880f442 ("net: Allow GRO to use and set levels of checksum unnecessary")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Jesse Brandeburg <jesse.brandeburg@intel.com>
Cc: Tom Herbert <tom@herbertland.com>
Acked-by: Willem de Bruijn <willemb@google.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/r/20210226212248.8300-1-daniel@iogearbox.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-02-28 12:01:08 -08:00
..
6lowpan treewide: replace '---help---' in Kconfig files with 'help' 2020-06-14 01:57:21 +09:00
9p net: remove redundant 'depends on NET' 2021-01-27 17:04:12 -08:00
802 net: 802: psnap.c: Use built-in RCU list checking 2020-02-24 13:02:53 -08:00
8021q Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net 2021-01-14 18:34:50 -08:00
appletalk appletalk: Fix skb allocation size in loopback case 2021-02-12 16:40:28 -08:00
atm net: atm: pppoatm: use new API for wakeup tasklet 2021-01-29 18:24:05 -08:00
ax25 Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net 2020-07-25 17:49:04 -07:00
batman-adv This feature/cleanup patchset is an updated version of the pull request 2021-02-08 11:32:40 -08:00
bluetooth Merge branch 'for-upstream' of git://git.kernel.org/pub/scm/linux/kern 2021-02-11 14:59:01 -08:00
bpf Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net 2021-01-20 12:16:11 -08:00
bpfilter net: remove redundant 'depends on NET' 2021-01-27 17:04:12 -08:00
bridge bridge: mrp: Update br_mrp to use new return values of br_mrp_switchdev 2021-02-16 14:47:46 -08:00
caif net: caif: Use netif_rx_any_context(). 2021-02-15 13:21:48 -08:00
can net: introduce CAN specific pointer in the struct net_device 2021-02-24 14:32:15 -08:00
ceph libceph: remove osdtimeout option entirely 2021-02-16 12:09:52 +01:00
core Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net 2021-02-16 17:51:13 -08:00
dcb net: dcb: use obj-$(CONFIG_DCB) form in net/Makefile 2021-01-27 17:03:52 -08:00
dccp dccp: Return the correct errno code 2021-02-06 11:15:28 -08:00
decnet net: decnet: fix netdev refcount leaking on error path 2021-01-27 17:33:46 -08:00
dns_resolver net: remove redundant 'depends on NET' 2021-01-27 17:04:12 -08:00
dsa net: dsa: tag_ocelot_8021q: fix driver dependency 2021-02-26 15:29:53 -08:00
ethernet net: datagram: fix some kernel-doc markups 2020-11-17 14:15:03 -08:00
ethtool ethtool: fix the check logic of at least one channel for RX/TX 2021-02-28 11:49:07 -08:00
hsr net: hsr: add support for EntryForgetTime 2021-02-25 09:41:51 -08:00
ieee802154 treewide: rename nla_strlcpy to nla_strscpy. 2020-11-16 08:08:54 -08:00
ife net: remove redundant 'depends on NET' 2021-01-27 17:04:12 -08:00
ipv4 net: Fix gro aggregation for udp encaps with zero csum 2021-02-28 12:01:08 -08:00
ipv6 net: icmp: pass zeroed opts from icmp{,v6}_ndo_send before sending 2021-02-23 11:29:52 -08:00
iucv net/af_iucv: build SG skbs for TRANS_HIPER sockets 2021-01-28 20:36:22 -08:00
kcm net: group skb_shinfo zerocopy related bits together. 2021-01-07 16:08:37 -08:00
key af_key: relax availability checks for skb size calculation 2021-01-04 10:05:50 +01:00
l2tp lsm,selinux: pass flowi_common instead of flowi to the LSM hooks 2020-11-23 18:36:21 -05:00
l3mdev net: l3mdev: use obj-$(CONFIG_NET_L3_MASTER_DEV) form in net/Makefile 2021-01-27 17:03:52 -08:00
lapb net: lapb: Copy the skb before sending a packet 2021-02-02 08:40:48 -08:00
llc net: remove redundant 'depends on NET' 2021-01-27 17:04:12 -08:00
mac80211 cfg80211/mac80211: Support disabling HE mode 2021-02-12 09:33:34 +01:00
mac802154 net: mac802154: convert tasklets to use new tasklet_setup() API 2020-11-07 10:40:56 -08:00
mpls mpls: drop skb's dst in mpls_forward() 2020-11-03 12:55:53 -08:00
mptcp mptcp: do not wakeup listener for MPJ subflows 2021-02-22 18:54:59 -08:00
ncsi net/ncsi: Use real net-device for response handler 2020-12-23 12:22:23 -08:00
netfilter netfilter: nftables: introduce table ownership 2021-02-15 18:17:15 +01:00
netlabel Merge https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net 2020-11-19 19:08:46 -08:00
netlink mptcp: avoid lock_fast usage in accept path 2021-02-12 16:31:46 -08:00
netrom treewide: Use fallthrough pseudo-keyword 2020-08-23 17:36:59 -05:00
nfc TTY/Serial driver changes for 5.12-rc1 2021-02-20 21:28:04 -08:00
nsh treewide: replace '---help---' in Kconfig files with 'help' 2020-06-14 01:57:21 +09:00
openvswitch net: openvswitch: add log message for error case 2021-01-14 16:32:14 -08:00
packet net/packet: Improve the comment about LL header visibility criteria 2021-02-06 14:59:28 -08:00
phonet treewide: Use fallthrough pseudo-keyword 2020-08-23 17:36:59 -05:00
psample net: psample: Fix netlink skb length with tunnel info 2021-02-25 09:49:46 -08:00
qrtr net: qrtr: Fix memory leak in qrtr_tun_open 2021-02-23 15:38:22 -08:00
rds net/rds: restrict iovecs length for RDS_CMSG_RDMA_ARGS 2021-02-02 08:44:08 -08:00
rfkill rfkill: add a reason to the HW rfkill state 2020-12-11 12:47:17 +01:00
rose rose: Fix Null pointer dereference in rose_send_frame() 2020-11-20 10:04:58 -08:00
rxrpc rxrpc: Fix dependency on IPv6 in udp tunnel config 2021-02-12 16:42:05 -08:00
sched net/sched: cls_flower: validate ct_state for invalid and reply flags 2021-02-23 12:23:44 -08:00
sctp Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net 2021-02-10 13:30:12 -08:00
smc net/smc: use memcpy instead of snprintf to avoid out of bounds read 2021-01-12 20:22:01 -08:00
strparser
sunrpc Optimization: 2021-02-22 13:29:55 -08:00
switchdev net: bridge: propagate extack through switchdev_port_attr_set 2021-02-14 17:38:11 -08:00
tipc tipc: remove duplicated code in tipc_msg_create 2021-01-27 18:50:07 -08:00
tls net/tls: Select SOCK_RX_QUEUE_MAPPING from TLS_DEVICE 2021-02-11 19:08:06 -08:00
unix af_unix: handle idmapped mounts 2021-01-24 14:27:18 +01:00
vmw_vsock vsock: fix locking in vsock_shutdown() 2021-02-09 15:31:22 -08:00
wireless cfg80211/mac80211: Support disabling HE mode 2021-02-12 09:33:34 +01:00
x25 net: x25: Remove unimplemented X.25-over-LLC code stubs 2020-12-12 17:15:33 -08:00
xdp xsk: Fold xp_assign_dev and __xp_assign_dev 2021-01-25 23:56:33 +01:00
xfrm Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/klassert/ipsec-next 2021-02-09 11:23:41 -08:00
compat.c iov_iter: transparently handle compat iovecs in import_iovec 2020-10-03 00:02:13 -04:00
devres.c net: devres: rename the release callback of devm_register_netdev() 2020-06-30 15:57:34 -07:00
Kconfig net/sock: Add kernel config SOCK_RX_QUEUE_MAPPING 2021-02-11 19:08:06 -08:00
Makefile net: l3mdev: use obj-$(CONFIG_NET_L3_MASTER_DEV) form in net/Makefile 2021-01-27 17:03:52 -08:00
socket.c idmapped-mounts-v5.12 2021-02-23 13:39:45 -08:00
sysctl_net.c