Is there any value to supporting fragments in a driver which
doesn't do hardware checksumming? IIRC Alexey had a patch to do
such for Tulip, but I don't see it in the above patchset.
I'm actually considering making the SG w/o hwcsum situation illegal.
Basically, with SG only, you can run into some problems.
We don't prevent anyone from making modifications to the
paged data, we just grab references to the pages.
This works perfectly fine when the card does the checksumming,
the card DMAs a snapshot of the data into it's internal buffers,
checksums that local snapshot of the data, and the checksum is fine.
If, on the other hand, we're doing this in software, we can send out
packets with bad checksums. The next retransmit could fix it up, but
this is a horrible scheme quality of implementation wise.
Later,
David S. Miller
davem@redhat.com
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/