On Mon, 2015-10-05 at 09:23 -0700, Tom Herbert wrote:

1) Drivers may advertise NETIF_F_HW_CSUM. The stack will indicate
checksum offload exclusively using the
CHECKSUM_PARTIAL/csum_start/csum_offset interface. No additional
interfaces (bits in skbuff should not be needed)
2) A driver may inspect packets via ndo_check to decide if it wants to
offload the checksum, if not cancels NETIF_F_HW_CSUM in the packet.
3) In driver xmit when CHECKSUM_PARTIAL is set the driver MUST
correctly resolve the checksum-- either by properly offloading to the
device or calling skb_checksum_help.

In cases where they haven't used .ndo_features_check() to ensure that
they don't *see* such packets, sure. But using .ndo_features_check()
should probably be the preferred method.

4) To help drivers for devices with limited offload capabilities we'll
define a helper function to check for typical restrictions (.e.g. IPv4
only, TCP/UDP only. no encapsulation, no IPv6 extension headers,
etc.). I am working on this helper function and will send RFC shortly.

I do suspect that helper function would benefit from seeing TCP/UDP
flags in the high bits of csum_offset, rather than grubbing around in
the packet for itself to see if it's really TCP/UDP. After all, it's
almost free to set those in the first place — at least for locally
-generated packets. And not *so* hard to add them in
skb_partial_csum_set(). But hey, if you can push an implementation
which is grubbing around in the packet then at least *I* don't have to
feel dirty for it... :)

-- 
David Woodhouse                            Open Source Technology Centre
David.Woodhouse@intel.com                              Intel Corporation