I sent a V3 since an unrelated change was also included in this patch.
Darrell
On Mon, Jan 28, 2019 at 10:41 AM Darrell Ball wrote:
>
>
> On Mon, Jan 28, 2019 at 10:05 AM Darrell Ball wrote:
>
>> 'conn_key_extract()' in OVS is including L2 (Ethernet) pad bytes
>>
>
> Above 'in OVS' was meant
On Mon, Jan 28, 2019 at 10:05 AM Darrell Ball wrote:
> 'conn_key_extract()' in OVS is including L2 (Ethernet) pad bytes
>
Above 'in OVS' was meant to be 'in userspace conntrack' - will fix later
if there are objections.
> for both L3 and L4 sizes. One problem is any packet with non-zero
> L2
I resent this patch with a missing co-author added.
Thanks Darrell
On Mon, Jan 28, 2019 at 10:05 AM Darrell Ball wrote:
> 'conn_key_extract()' in OVS is including L2 (Ethernet) pad bytes
> for both L3 and L4 sizes. One problem is any packet with non-zero
> L2 padding can incorrectly fail L4 che
'conn_key_extract()' in OVS is including L2 (Ethernet) pad bytes
for both L3 and L4 sizes. One problem is any packet with non-zero
L2 padding can incorrectly fail L4 checksum validation.
This patch fixes conn_key_extract() by ignoring L2 pad bytes.
Fixes: a489b16854b5 ("conntrack: New userspace c
'conn_key_extract()' in OVS is including L2 (Ethernet) pad bytes
for both L3 and L4 sizes. One problem is any packet with non-zero
L2 padding can incorrectly fail L4 checksum validation.
This patch fixes conn_key_extract() by ignoring L2 pad bytes.
Fixes: a489b16854b5 ("conntrack: New userspace c