Message ID | 20240212081344.158817-2-mkp@redhat.com |
---|---|
State | Superseded |
Headers | show |
Series | [ovs-dev,1/4] Userspace: Software fallback for UDP encapsulated TCP segmentation. | expand |
Context | Check | Description |
---|---|---|
ovsrobot/apply-robot | warning | apply and check: warning |
ovsrobot/github-robot-_Build_and_Test | fail | github build: failed |
ovsrobot/intel-ovs-compilation | success | test: success |
Bleep bloop. Greetings Mike Pattrick, I am a robot and I have tried out your patch. Thanks for your contribution. I encountered some error that I wasn't expecting. See the details below. checkpatch: WARNING: The subject summary should end with a dot. Subject: netdev-linux: Favour inner packet for multi-encapsulated tso Lines checked: 54, Warnings: 1, Errors: 0 Please check this out. If you feel there has been an error, please email aconole@redhat.com Thanks, 0-day Robot
diff --git a/lib/netdev-linux.c b/lib/netdev-linux.c index 1b2e5b6c2..7a156cc28 100644 --- a/lib/netdev-linux.c +++ b/lib/netdev-linux.c @@ -7239,14 +7239,23 @@ netdev_linux_prepend_vnet_hdr(struct dp_packet *b, int mtu) vnet->csum_offset = (OVS_FORCE __virtio16) __builtin_offsetof( struct tcp_header, tcp_csum); } else if (dp_packet_hwol_l4_is_udp(b)) { - struct udp_header *udp_hdr = dp_packet_l4(b); + /* Favour the inner packet when indicating checksum offsets. */ + void *l3_off = dp_packet_inner_l3(b); + void *l4_off = dp_packet_inner_l4(b); + + if (!l3_off || !l4_off) { + l3_off = dp_packet_l3(b); + l4_off = dp_packet_l4(b); + } + struct udp_header *udp_hdr = l4_off; + ovs_be16 csum = 0; if (dp_packet_hwol_is_ipv4(b)) { - const struct ip_header *ip_hdr = dp_packet_l3(b); + const struct ip_header *ip_hdr = l3_off; csum = ~csum_finish(packet_csum_pseudoheader(ip_hdr)); } else if (dp_packet_hwol_tx_ipv6(b)) { - const struct ovs_16aligned_ip6_hdr *ip6_hdr = dp_packet_l3(b); + const struct ovs_16aligned_ip6_hdr *ip6_hdr = l4_off; csum = ~csum_finish(packet_csum_pseudoheader6(ip6_hdr)); }
Previously if an OVS configuration nested multiple layers of UDP tunnels like VXLAN or GENEVE ontop of each other through netdev-linux interfaces, the vnet header would be incorrectly set to the outermost UDP tunnel layer instead of the intermediary tunnel layer. This resulted in the middle UDP tunnel not checksum offloading properly. Fixes: 3337e6d91c5b ("userspace: Enable L4 checksum offloading by default.") Reported-by: David Marchand <david.marchand@redhat.com> Signed-off-by: Mike Pattrick <mkp@redhat.com> --- lib/netdev-linux.c | 15 ++++++++++++--- 1 file changed, 12 insertions(+), 3 deletions(-)