On 10/12/26 11:50 AM, Sahaj Chaudhari wrote: > Send the crafted later fragment through OVS_PACKET_CMD_EXECUTE to preserve > the unset transport-header path from the report. Set frag=2 in the action key > and match DEBUG_NET warnings at the skb header check rather than a > potentially inlined helper. Skip when CONFIG_DEBUG_NET cannot be confirmed > and enable CONFIG_DEBUG_KERNEL in the OVS fragment. Drop the flow-counter > assertion because its SET-action dump relies on broken formatting. > > Link: https://lore.kernel.org/r/[email protected] > Signed-off-by: Sahaj Chaudhari <[email protected]> > --- > .../testing/selftests/net/openvswitch/config | 1 + > .../selftests/net/openvswitch/openvswitch.sh | 33 ++++++++------- > .../selftests/net/openvswitch/ovs-dpctl.py | 40 +++++++++++++++++++ > .../net/openvswitch/ovs-ipv6-frag.py | 27 ++++++++----- > 4 files changed, 76 insertions(+), 25 deletions(-) Hi, Sahaj. This patch is not formatted properly. Please, read the documentation on how to format and submit patches before sending. Also, every new version must be a complete patch, not a diff with the previous one, which I believe, is also covered by the submitting patches guide in the docs.
For the patch itself, I don't think it is worth adding 100 lines of test code for this specific issue. If we do that for every one-liner fix, the kernel code will just explode in size. If you have some thoughts on how to make a more generic test that would be useful not just for this one issue, but would cover a larger class of functionality that is not already covered, then we could consider that. But as it is right now, I don't think we need such a narrow test. Best regards, Ilya Maximets. _______________________________________________ dev mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-dev
