[UPDATE] Zerocopy patches, against 2.4.1-pre10

[David S. Miller]

I'm back from OZ, and to help deal with my sudden lack of Victoria
Bitter, I've made a new zerocopy patch set.  You will notice that
it is now significantly smaller than previous versions.  This is
because all of the straight bug fixes and cleanups in my tree made
it into 2.4.1-pre10.  What remains is purely zerocopy and nothing
more.

ftp://ftp.kernel.org/pub/linux/kernel/people/davem/zerocopy-2.4.1p10-1.diff.gz

If you don't see it immediately, give it a few minutes to propagate.
Thanks.

I have several bug fixes (from Paul Russell, and Alexey) and updates
(from the DecNET folks) in my queue, so if you don't see your favorite
change don't worry just yet.

There is one critical failure I saw reported with zerocopy, where all
transmits basically failed using a 3c59x card.  This indicates that
our driver checks thought the 3c59x you had supported TX checksumming
in hardware, when in fact it does not.  Please, if you get this
behavior, provide the "dmesg" logs with the 3c59x driver messages,
also please provide /proc/pci output.  Both things should assist us
in determining what we are doing wrong with TX csum capability
detection.

Basically, this is a show-stopper bug in the zerocopy patches and
must be fixed.

Later,
David S. Miller
[e-mail]