On Thu, 2017-01-26 at 06:45 -0800, Eric Dumazet wrote:

On Thu, 2017-01-26 at 14:49 +0100, Johannes Berg wrote:

quoted

Oops, sorry - receive. We can only indicate "CHECKSUM_UNNECESSARY",
nothing more advanced right now, but right now we'd indicate that
if
the packet had 0x0000 in the checksum field, but should've had
0xffff.

On TX I believe we actually do in HW exactly what your patch just
did.

Can you describe the visible effects of this problem ?

Is that because of a conversion we might do later to
CHECKSUM_COMPLETE ?

Unfortunately, I haven't been able to actually test this yet. I also
didn't find the code that would drop frames with CSUM 0 either, so I'm
thinking - for now - that if all the csum handling is skipped, dropping
0 csum frames would also be, and then we'd accept a frame we should
actually have dropped.

I'll go test this I guess :)

Any pointers to where 0 csum frames are dropped?

johannes