bugfix patch to arch/ppc/8260_io/fcc_enet.c

[Dan Malek]

David Schleef wrote:

> The other (and more likely) is:

>         if(cep->skb_cur == cep->skb_dirty)break;

OK, I kind of found this too.  My last message indicated this
should have been fixed, and if this _exact_ line of code is in
your driver, it is too old.  The proper test checks for the buffer
pointers equal, and also adds the 'tx_full' condition so the
wrap around case is properly handled.

The current version of this driver has been subjected to the
University of New Hampshire Interoperability testing lab by
a very large communication company.  The test reports were sent
to me, I made some changes according to their recommendations,
and I believe they were happy with the results.  I would hope
they found all of the errors :-).


	-- Dan

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/